No character whose development is dependent in greater or less degree on

the stimulation of some substance derived from the gonads can have

originated as a mutation, because the term mutation means a new character

which develops in the soma as a result of the loss or gain of some factor

or determinant in the chromosomes. To say that certain mutations consist

of new factors which only the development of characters in the soma when

the part of the soma concerned is stimulated by a hormone, is a mere

assertion unsupported at present by any evidence. As an example of the way

in which Mendelians misunderstand the problem to be considered, I may

refer to Doncaster's book, _The Determination of Sex_ [Footnote: Camb.

Univ. 1914, p. 99.] in which he remarks: 'It follows that the secondary

sexual characters cannot arise simply from the action of hormones; they

must be due to differences in the tissues of the body, and the activity of

the ovary or testis must be regarded rather as a stimulus to their

development than as their source of origin.' This seems to imply a serious

misunderstanding of the idea of the action of the hormones from the gonads

and of hormones in general. No one would suggest that the hormones from

the testis should be regarded as in any sense the origin of the antlers of

a stag. If so, why should not antlers equally develop in the stallion or

in the buck rabbit, or indeed in man? How far Doncaster is right in

holding that the soma is different in the two sexes is a question already

mentioned, but it is obvious that in each individual the somatic sexual

characters proper to its species are present potentially in its

constitution by heredity--in other words, as factors or determinants in

the chromosomes of the zygote from which it was developed; but the normal

development of such characters in the individual soma is either entirely

dependent on the stimulus of the hormone of the gonad or is profoundly

influenced by the presence or absence of that stimulus. The evidence, as

we have seen, proves that, at any rate in the large number of cases where

this relation between somatic sex-characters and hormones produced by the

reproductive organs exists, the characters are inherited by both sexes. In

one sex they are fully developed, in the other rudimentary or wanting. But

the sex, usually the female, in which they are rudimentary or wanting is

capable of transmitting them to offspring, and also is capable of

developing them more or less completely when the ovaries are removed,

atrophied or diseased. If we state these facts in the terms of our present

conceptions of chromosomes and determinants or factors, we must say that

the factors for these characters are present in the chromosomes of both

male and female gametes. The question then is, how did these factors

arise? If they were mutations not caused by any influence from the

exterior, what is the reason why these particular characters which alone

have an adaptive relation to the sexual or reproductive habits of the

animal are also the only characters which are influenced by the hormones

of the reproductive organs? The idea of mutations implies neither an

external relation nor an internal relation in the organ or character; but

these characters have both, the external relation in the function they

perform in the sexual life of the individual, the internal relation in the

fact that their development is affected by the sexual hormones. There is

no more striking example of the inadequacy of the current conceptions of

Mendelism and mutation to cover the of bionomics and evolution.

The truth is that facts and experiments within a somewhat narrow field

have assumed too much importance in recent biological research. No

increase in the number of facts or experimental results of a particular

class will compensate for the want of sound reasoning and a comprehensive

grasp of the phenomena to be explained. The coexistence of the external

and the internal relation in the characters we are considering suggests

that one is the cause of the other, and as it is obvious that the relation

for instance of a stag's antlers to a testicular hormone could not very

well be the cause of the use of the antlers in fighting, the reasonable

suggestion is that the latter is the cause of the former. We have already

seen that the development and shedding of the antler are processes of

essentially the same kind physiologically, or pathologically, as these

which can be and are occasionally produced in the individual soma by

mechanical stimulus and injury to the periosteum. The fact that a hormone

from the testis affects the development of the antler, as well as our

knowledge of hormones in general, suggests a special theory of the

heredity of somatic modifications due to external stimuli. Physiologists

are apt to look for a particular gland to produce every internal

secretion. But the fact that the wall of the intestine produces secretion,

which carried by the blood causes the pancreas to secrete, shows that a

particular gland is not necessary. There is nothing improbable in

supposing that a tissue stimulated to excessive growth by external

irritation would give off special substances to the blood. We know that

living tissues give off products, and that these are not merely pure CO2

and H2O, but complicated compounds. The theory proposed by me in 1908 was

that we have within the gonads numerous gametocytes whose chromosomes

contain factors corresponding to the different parts of the soma, and that

factors or determinants might be stimulated by products circulating in the

blood and derived from the parts of the soma corresponding to them. There

is no reason to suppose that an exostosis formed on the frontal bone as a

result of repeated mechanical stimulation due to the butting of stags

would give off a special hormone which was never formed in the body

before, but it would probably in its increased growth give off an

increased quantity of intermediate waste products of the same kind as the

tissues from which it arose gave off before. These products would act as a

hormone on the gametocytes, stimulating the factors which in the next

generation would control the development of the frontal bone and adjacent

tissues.

The difficulty of this theory is one which has occurred to biologists who

have previously made suggestions of a connexion between hormones and

heredity--namely, how hormones or waste products from one part of the body

could differ from these from the same tissue in another part of the body.

If there were no special relation, hypertrophy of bone on one part of the

body such as the head, would merely stimulate the factor for the whole

skeleton in the gametocytes, and the result would merely be an increased

development of the whole skeleton. On the other hand, we have the evident

fact that a number of chromosomes formed apparently of the same substance,

by a series of equal chromosome divisions determine all the various

special parts of the complicated body. This is not more difficult to

understand than that every part of the body should give off special

substances which would have a special effect on the corresponding parts of

the chromosomes. We know that skin glands in different parts of the body

produce special odours, although all formed of the same tissue and all

derived from the epidermis. It seems not impossible that bones of

different parts of the body give off different hormones. If the factors in

the gametes were thus stimulated they would, when they developed in a new

individual, product a slightly increased development of the part which was

hypertrophied in the parent soma. No matter how slight the degree of

hereditary effect, if the stimulation was repeated in every generation, as

in the case of such characters as we are considering it undoubtedly was,

the hereditary effect would constantly increase until it was far greater

than the direct effect of the stimulation. We may express the process

mathematically in this way. Suppose the amount of hypertrophy in such a

case as the antlers to be _x,_ and that some fraction of this is

inherited. Then in the second generation the same amount of stimulation

together with the inherited effect would produce a result equal to

_x+x/n_. The latter fraction being already hereditary, a new fraction

_x/n_ would be added to the heredity in each generation, so that after _m_

generations the amount of hereditary development would be _x+mx/n_. If _n_

were 1000, then after 1000 generations the inherited effect would be equal

to _x_. This, it is true, would not be a very rapid increase. But it is

possible that the fraction _x/n_ would increase, for the heredity might

very well consist not only in a growth independent of stimulation, but in

an increasing response to stimulation, so that _x_ itself might be

increasing, and the fraction _x/n_ would become larger in each generation.

The death and loss of the skin over the antler, originally duo to the

laceration of the skin in fighting, has also become hereditary, and it is

certainly difficult to conceive the action of hormones in this part of the

process. All we can suggest is that the hormone from the rapidly growing

antler, including the covering skin, is acting on the corresponding factor

in the gametocytes for a certain part of every year, and then, when the

skin is stripped off, the hormone disappears. The factor then may be said

to be stimulated for a time and then the stimulus suddenly ceases. The

bone also begins to die when the skin and periosteum is stripped off, and

the hormone from this also ceases to be produced.

The annual shedding and recrescence of the antler, however, is only to be

understood in connexion with the effect of the testicular hormone.

According to my theory there are two hormone actions, the centripetal from

the hypertrophied tissue to the corresponding factor in the gametocytes,

and the centrifugal from the testis to the tissue of the antler or other

organ concerned. The reason why the somatic sexual character does not

develop until the time of puberty, and develops again each breeding season

in such cases as antlers, is that the original hypertrophy due to external

stimulation occurred only when the testicular hormone was circulating in

the blood. The factor in the gametocytes then in each generation acted

upon by both hormones, and we must suppose that in some way the result was

produced that the hereditary development of the antler in the soma only

took place when the testicular hormone was present. It is to be remembered

that we are unable at present to form a clear conception of the process

of development, to understand how the simple fertilised ovum is able by

cell-division and differentiation to develop into a complicated organism

with organs and characters predetermined in the single cell which

constitutes the ovum. If we accept the idea that characters are

represented by particular parts of the chromosomes, according to Morgan's

scheme, our theory of development is the modern form of the theory of

preformation. When in the course of development the cells of the head from

which the antlers arise are formed, each of these cells must be supposed

to contain the same chromosomes as the original ovum from which the cells

have descended by repeated cell-division. The factors in these chromosomes

corresponding to the forehead have been stimulated while in the parent

animal by hormones from the outgrowth of tissue produced by external

mechanical stimulation, while at the same time they were permeated by the

testicular hormone produced either by the gametocytes themselves or by

interstitial cells of the testis. When the head begins to form in the

process of individual development, the factors, according to my theory,

have a tendency to form the special growth of tissue of which the

incipient antler consists, but part of the stimulus is wanting, and is not

completed until the testicular hormone is produced and diffused into the

circulation--that is to say, when the testes are becoming mature and

functional.

I do not claim that this theory in complete--it is impossible to

understand the process completely in the present state of knowledge--but I

maintain that it is the only theory which affords any explanation of the

remarkable facts concerning the influence of the hormones from the

reproductive organs on the development of secondary sexual characters,

while at the same time explaining the adaptive relation of these

characters or organs to the sexual habits of the various species. On the

mutation hypothesis, adaptation is purely accidental. T. H. Morgan

considers that the appearance of two slightly different shades of eye

colour in male and female in a culture of a fruit-fly in a bottle is

sufficient to settle the whole problem of sexual dimorphism, and to

supersede Darwin's complicated theory of sexual selection. The possibility

of a Lamarckian explanation he does not even mention. He would doubtless

assume that the antlers of stags arose as a mutation, without explaining

how they came to be affected by the testicular hormone, and that when they

arose the stags found them convenient as fighting weapons. But the

complicated adaptive relations are not to be disposed of by the simple

word mutation. The males have sexual instincts, themselves dependent on

the testicular hormone, which develop sexual jealousy and rivalry, and the

Ruminants fight by butting with their heads because they have no incisor

teeth in the upper jaw, or tusks, which are used in fighting in other

species. Doubtless, mutations have occurred in antlers as in other

characters; in fact all hereditary characters are subject to mutation.

This in the most probable explanation, not only of the occasional

occurrence of hornless individual stags, but of the differences between

the antlers of different species, for there is no reason to believe that

the special character of the antler in each species is adapted to a

special mode of fighting in each species.

The different structure of the horns of the Bovine and Ovine Ruminants is,

in my view, the result of a different mode of fighting. If we suppose that

the fighting was slower and less fierce in the Bovidae, so that the skin

over the exostosis was subject to friction but not lacerated, the result

would be a thickening of the horny layer of the epidermis as we find it,

and the fact that the skin and periosteum are not destroyed explains why

the horns are not shed but permanent.

There is a tendency among Mendelians and mutationists to overestimate the

importance of experiments in comparison with reasoning, either inductive

or deductive. Bateson, however, has admitted that Mendelian experiments

and observations on mutation have not solved the problem of adaptation. It

seems to be demanded, nevertheless, that characters must be produced

experimentally and then inherited before the hereditary influence of

external stimuli can be accepted. Kammerer's experiments in this direction

have been sceptically criticised, and it must be granted that the evidence

he has published is not sufficient to produce complete conviction. But

experiments of this kind are from the nature of the case difficult if not

impossible. There is, however, another method--namely, to take a character

which is certainly to some extent hereditary, and then to ascertain by

experiment if it is 'acquired.' If it be proved that a hereditary

character was originally somatogenic, it follows that somatogenic

characters in time become hereditary. This is the reasoning I have used in

reference to my experiments on the production of pigment on the lower

sides of Flat-fishes, and I obtained similar evidence with regard to the

excessive growth of the tail feathers in the Japanese Tosa-fowls,

[Footnote: 'Observations and Experiments on Japanese Long-tailed Fowls,'

_Proc. Zool. Soc._, 1903.] which is a modification of a secondary sexual

character. In these fowls the feathers of the tail in the hens are only

slightly lengthened.

I learned from Mr. John Sparks, who himself brought specimens of the breed

from Japan, that the Japanese not only keep the birds separately on high

perches in special cages, but pull the tail feathers gently every morning

in order to cause them to grow longer. One question which I had to

investigate on my specimens, hatched from eggs obtained from Mr. Sparks,

was the relation of the growth of the feathers to the moult which occurs

in ordinary birds. My experiment consisted in keeping two cocks, A and B,

the first of which was left to itself, while in the second the feathers

were gently pulled by stroking between the finger and thumb from the base

outwards. The feathers in the tail were seven pairs of rectrices, two rows

of tail coverts, anterior and posterior, four or five pairs in each row, a

number of transition feathers: all these were steel-blue, almost black; in

front of them on the saddle were a number of reddish yellow, very slender

saddle hackles.

In September 1901, when the birds ware just over three months old, the

adult feathers of the tail were all growing. The growing condition can be

distinguished by the presence of a horny tubular sheath extending up the

base of the feather for about one inch. When growth ceases this sheath is

shed. In cock A growth continued till the end of the following March, when

the longest feathers, the central rectrices, 2 feet 4-1/2 inches long. One

of the feathers--namely, one of the anterior tail coverts--was

accidentally pulled out on 11th February 1902, when it was 15-1/4 inches

long and had nearly ceased to grow and formed its quill, and it

immediately began to grow again and continued to grow till the following

September, when it was accidentally broken off at the base: it was then 18

inches (44.5 cm.) long.

The effect of stroking in cock B was to pull out from time to time one of

the growing feathers. Of the original feathers, one, the left central

posterior covert, continued to grow till 13th July 1902, when it was 2

feet 9-1/2 inches long without the part contained in the follicle. All the

feathers pulled out immediately commenced to grow again, except the last

two pulled out 27th May and 13th July, which did not grow again till the

following moulting season, in September.

The first right central rectrix in cock B was accidentally pulled out on

13th April 1902, when it was 2 feet 9-7/8 inches long. Its successor began

to grow immediately, and in course of time pieces of it were broken off

accidentally without injury to the base in the socket, which continued to

grow until 16th June 1905, when it torn out of its socket. The total

length of the feather with the pieces previously broken off, which were

measured and preserved, was 11 feet 5-1/2 inches. It therefore continued

to grow without interruption for three years and two months at an average

rate of 3.6 inches per month.

In cock A only four of the short outer rectrices were moulted in the

beginning of September 1902: the longer feathers--namely, central

rectrices and tail coverts--which ceased to grow naturally in the spring

of 1902, were not moulted till the beginning of October. This shows the

great importance of pulling out the feathers as soon as they show signs of

ceasing to grow, in order to obtain the abnormally long feathers. The

central rectrices continued to grow till the beginning of September 1903,

when that of the left side was 3 feet 6 inches long, that of the right

about an inch shorter. The coverts had ceased to grow of their own accord

some time before this, and the central ones of the posterior row were

about 3 feet long.

As it seemed possible that there was some natural congenital difference in

growth of feathers between cocks A and B, I commenced early in March 1903

to pull and stroke the feathers of the left side only in cock A, leaving

those of the right side untouched. On 30th July on the left side the

central rectrix and the first and second posterior coverts were still

growing, on the right side the central rectrix was also growing, but the

first and second posterior coverts had ceased growth and formed their

quills. The first posterior covert on the left or pulled side was 3 inches

longer than that of the right. The second posterior covert on the left

side was still longer. The first and second posterior coverts of left side

did not cease growth till 26th August. On 2nd September the left central

rectrix was almost at the end of its growth, the right had ceased to grow

a little before. The left was about an inch longer than the right. Thus

both in length in duration of growth the feathers of the pulled side were

longer than those of the right, and this was the result of treatment

continued only six months, and commenced some months after the feathers

had begun to grow. I have no doubt, however, that the pulling out of the

feather as soon as it shows signs of forming quill, so that its successor

at once grows again, is even more important in producing the great length

of feather than the stroking of the feather itself.

In this case, then there is no doubt (_a_) that the long-tailed birds are

artificially treated with the utmost care and ingenuity by the Japanese,

who produced them; (_b_) that the mechanical stimulus in my experiments

did cause the feathers to grow for a longer period and attain greater

length; (_c_) that the tendency to longer growth is, even when no

treatment is applied, distinctly inherited. It is a legitimate and logical

conclusion that the inherited tendency is the result of the artificial

treatment. No other breed of fowls shows such excessive growth of tail

feathers. It may be admitted that individuals differ considerably in their

congenital tendency to greater growth, _i.e._ greater length of the tail

feathers, but according to my views this is not contradictory to the main

conclusion, for every hereditary character shows individual variation.

It may be pointed out here that on the Lamarckian theory the conception of

adaptations is not teleological: they do not exist for a certain purpose,

but are the result of external stimulations arising from the actions and

habits of the organism. The latter conception is the more general, for

cases of somatic sexual characters exist which cannot be said to have a

use or function. For example, the comb and wattles of _Gallus_ are

sexually dimorphic, being in the original species larger in the cock than

in the hen. There is no convincing evidence that these appendages are

either for use or ornament. They are, in fact, a disadvantage to the bird,

being used by his adversary to take hold of when he strikes. The first

thing that happens when cocks fight is the bleeding and laceration of the

comb, as they peck at each other's heads. This laceration of the skin is,

in my view, the primary cause of the evolution of these structures,

leading to hypertrophy. But in this, as in other cases, the hereditary

result is regular, constant, and symmetrical, while the immediate effect

on the individual is doubtless irregular.

CHAPTER V

Mammalian Sexual Characters

Evidence Opposed To The Hormone Theory

Perhaps the most remarkable of all somatic sexual characters are those

which are almost universal in the whole class of Mammalia, the mammary

glands in the female, the scrotum in the male. We have considered the

evidence concerning the relation of the development and functional action

of the milk glands to hormones arising in the ovary or uterus, now we have

to consider the origin of the glands and of their peculiar physiology in

evolution. The obvious explanation from the Lamarckian point of view, and

in my opinion the true one, is that they owed their origin at the

beginning to the same stimulation which is applied to them now in every

female mammal that bears young. There is, as we have seen, a difficulty in

explaining how the occurrence of parturition causes the secretion of milk

to begin, but it is certain that the secretion soon stops if the milk is

not drawn from the glands by the sucking action of the offspring, or the

artificial imitation of that action. A cow that is not milked or milked

incompletely ceases to give milk. When the stimulus ceases, lactation

ceases. The pressure of the secretion in the alveoli causes the cells to

cease to secrete, much in the same way that pressure in the ureters

injures the secretory action of the renal epithelium. In the earliest

Mammals we may suppose that the young were born in a well-developed

condition, for at first the supply of milk would not have been enough to

sustain them for a long time as their only food. We must also suppose that

the mother began to cherish the young, keeping them in contact with her

abdomen. Then being hungry they began to suck at her hair or fur. The

actual development of the milk glands in Marsupials has been described by

Bresslau [Footnote: Stuttgart, 1901.] and by O'Donoghue. [Footnote:

_Q.J.M.S._, lvii., 1911-12.] The rudiment of the teat is a depression or

invagination of the epidermis from the bottom of which six stout hairs

arise. The follicles of these hairs extend down into the derma, and from

the upper end of the follicle, _i.e._ near the aperture of the

invagination, a long cellular outgrowth extends down into the derma,

branches at its end, and becomes hollow. These branches are the tubules of

the future milk gland. Another outgrowth from the follicle forms a

sebaceous gland. Later on the hairs and the sebaceous glands entirely

disappear, and the milk gland alone is left with its tubules and ducts

opening into the cavity of the teat. This is clear evidence that the milk

gland was evolved in connexion with hairs, and was an enlargement of

glands opening into the hair follicle, but it is difficult to understand

why a sebaceous gland is developed and afterwards disappears. This would

seem to indicate that the milk gland was not a hypertrophied sebaceous

gland, but a distinct outgrowth, which however had nothing to do with

sweat glands.

That the intra-uterine gestation, or its cessation, were not originally

necessary to determine the functional periodicity of the milk glands is

proved by their presence in the Monotremes, which are oviparous. It is

evident from the conditions in these mammals that both hair and milk

glands were evolved before the placenta.

It may also be pointed out here that, according to the evidence of

Steinach, in the milk glands at least among somatic sexual characters

there is no difference between the male and female in the heredity of the

organs. The zygote therefore, whether the sex of it is determined as male

or female, has the same factor for the development of milk glands. On the

chromosome theory as formulated by Morgan this factor must be in the

somatic chromosomes and not in the sex-chromosomes, and must be present in

every zygote. All the cells of the body, assuming that somatic segregation

does not occur, must possess the same chromosomes as the zygote from which

it developed, and whether the sex chromosomes are _XX_ or _XY_ or _X_,

there must be at any rate one chromosome bearing the factor for milk

glands. The functional development of these depends normally, according

to the evidence hitherto discovered, on the presence or absence of

hormones from the ovary or from the uterus.

If we attribute, as in my opinion we must, the primary origin of the milk

glands in evolution to the mechanical stimulus of sucking, we may attempt

to reconstruct the stages of the evolution of the present relation of the

glands to the other organs and processes of reproduction. In the earliest

stage represented by the Monotremata or Prototheria, there was no

intra-uterine development. We must suppose that in the beginning the

sucking stimulus caused both growth and secretion, for at first there was

nothing but sebaceous or sweat glands, and although a mutation might be

supposed to have produced larger glands, no mutation could explain the

influence of hormones on the growth and function of such glands. Then

heredity of the effect of stimulus took place to some slight degree, and

this would occur, according to my theory, only in the presence of the

hormone from the ovary in the same condition as that in which the

modification was first caused. This would be of course after ovulation,

and after hatching of the eggs. In the next stage, if we adopt the modern

view that Marsupials are descended from Placental Mammals, the eggs would

be retained for increasing periods in the uteri, and would be born in a

well-developed condition, since lactation would demand active sucking

effort on the part of the young. The early Placentalia would inherit from

the Monotreme-like ancestors the development of the milk glands after

ovulation, although no sucking was taking place while the young were

inside the uterus. It seems probable that the relation between parturition

and actual milk secretion originated with the sucking stimulus of the

young after birth.

There is good evidence that the secretion of milk may continue almost

indefinitely under the stimulus of sucking or milking. Neither

menstruation nor gestation put an end to it. Cows may continue to give

milk until the next parturition, and if castrated during lactation will

continue to yield milk for years. Women also may continue to produce milk

as long as the child is allowed to suck, and this has been in some cases

two or three years or even more. Moreover, lactation may be induced by the

repeated act of sucking without any gestation. This has happened in mares,

virgin bitches, mules, virgin women, and in one woman lactation continued

uninterruptedly for forty-seven years, to her eighty-first year, long

after the ovary had ceased to be functional. Lactation has also been

induced in male animals, _e.g._ in a bull, a male goat, male sheep, and in

men. [Footnote: Knott, 'Abnormal Lactation,' _American Medicine_, vol. ii

(new series), 1907.] We may conclude, therefore, that the secretion of

milk normally begins by heredity after parturition, and this, in

accordance with what we have learned about hormones in connexion with the

reproductive system, is probably the consequence of the withdrawal of the

hormone absorbed from the foetus. I do not think it is necessary to

suppose, as do Lane-Claypon and Starling, that the hormone physiologically

inhibits the dissimilative process and augments the assimilative, and that

the withdrawal of the hormone at parturition therefore causes the

dissimilative process, _i.e._ secretion of milk. My conclusion is that the

process of secretion set up by the mechanical stimulus of sucking is

inherited as it was acquired, so that it only begins to take place in the

individual in the absence of the hormone from the foetus, which was absent

when the process was acquired. The growth of the gland during gestation

would then be due to the postponement of the process of secretion in

consequence of the presence of the foetal hormone, and in this way this

hormone has become in the course of evolution at once the stimulus to

growth and the cause of the inhibition of secretion.

This interpretation does not, however, agree with the case of _Dasyurus_.

If the foetal hormone is absorbed from the pouch, as I have suggested, in

order to explain the persistence of the corpora lutea during lactation,

then the secretion of milk after parturition ought not to take place. But

in this case the sucking stimulus has been applied to the glands after a

very short gestation, while the hormone from the foetus is being absorbed

in the pouch, and therefore the hereditary correlation between secretion

and absence of foetal hormone may be assumed to have been lost in the

course of evolution.

We have next to consider the question of the evolution of the corpora

lutea. If these bodies are formed only in Mammals which have uterine

gestation, and not in Prototheria, they cannot be the only essential

source of the hormone which stimulates the development of the milk glands,

since the latter develop in Prototheria. Again it is difficult, it might

be said impossible, to believe that an accidental mutation gave rise to

corpora lutea the secretion of which caused uterine gestation and

ultimately the formation of the placenta. It seems more probable that the

retention of the originally yolked ova within the oviduct, however this

retention arose, was the essential cause of the formation of the placenta

and all the changes which the uterus undergoes in gestation. The

absorption of nutriment from the walls of the uterus, and the chemical and

mechanical stimulation of those walls, might well be the cause of the

diversion of nutrition from the ovary, leading gradually to the decline of

the process of secretion of yolk in the ova.

The conceptions and the mode of reasoning of the physiologist are very

different from those of the evolutionist. The former concludes from

certain experiments that a given organ of internal secretion has a certain

function. The corpora lutea, for example, according to one theory are

ductless glands, the function of whose secretion is to establish ova in

the uterus and promote their development. Another function suggested for

the secretion of the corpora lutea is to prevent further ovulation during

pregnancy. The evolutionist, on the other hand, asks what was the origin

of this corpora lutea, why should the ruptured ovarian follicles after the

escape of the ova in Mammals undergo a progressive development and persist

during the greater part of the whole of pregnancy? It seems obvious that

the corpora lutea in evolution were a consequence of intra-uterine

gestation, for they occur only in association with this condition, and it

is impossible to suppose that a mutation could arise accidentally by which

the ruptured follicles should produce a secretion which would cause the

fertilised ova to develop within the oviducts. The developing ovum within

the uterus may, however, reasonably be supposed to give off something

which is absorbed into the maternal blood, and this something would be of

the same nature as that which was given off by the ovum while still within

the ovarian follicle. The presence of this hormone might cause the

follicular cells to behave as though the ovum was still present in the

follicle, so that they would persist and not die and be absorbed. But this

leaves the question, what is lutein and why is it secreted? Lutein is a

colouring matter sometimes found in blood-clots, and probably derived from

haemoglobin. In the corpus luteum the lutein is contained in the cells,

not in a blood-clot.

Chemical investigation shows that the lutein of the corpus luteum is

almost if not quite identical with the colouring matter of the yolk in

birds and reptiles. Escher [Footnote: _Ztschr. f. Physiol. Chem._, 83

(1912).] found that the lutein of the corpus luteum had the formula

C{40}H{56} and was apparently identical with the carotin of the carrot,

while the lutein of egg-yolk was C{40}H{56}O{2} and more soluble in

alcohol, less soluble in petroleum ether, than that of the corpus luteum.

The difference, if it exists, is very slight, and it is evident that one

compound could easily be converted into the other. Moreover, the

hypertrophied follicular cells which constitute the corpus luteum secrete

fat which is seen in them in globules. The similarity of their contents

therefore to yolk is very remarkable, and it may be suggested that the

hormones absorbed from the ovum or embryo in the uterus acts upon the

follicular cells in such a way as to cause them to secrete substances

which in the ancestor were passed on to the ovum and formed the yolk. It

may be urged that this idea is contradictory to the previous suggestion

that the absorption of nourishment by the intra-uterine embryo was the

cause of the gradual decline of the process of yolk-secretion by the ova

in the ovary, but it is not really so. Originally in the reptilian

ancestor, or in the Monotreme, the ovum in the follicle secreted

yellow-coloured yolk. The materials for this, at any rate, passed through

the follicle cells, and it is probable that these cells were not entirely

passive, but actively secretory in the process. Substances diffusing from

the ovum would be present in the follicle cells during this process, and

probably act as a stimulus. The same substances diffusing from the ovum

during its development in the uterus would continue to stimulate the

follicle cells, and thus explain not merely their persistence, but their

secretory activity. The ovum being no longer present in the ovary, the

secretions would remain in the follicular cells, and the corpus luteum

would be explained.

If this theory is sound, it would follow that corpora lutea are not formed

in cases where the ova are not retained in the oviduct during their

development. The essential process in the development of these structures

is the hypertrophy and, in some cases at least, multiplication of the

follicular cells in the ruptured follicle. I have already mentioned that

this process does not occur in Teleosteans whose ovaries were studied by

me. These were species of Teleosteans in which fertilisation is external.

Marshall, in his _Physiology of Reproduction_, [Footnote: London, 1910, p.

151.] quotes a number of authors who have published observations on the

changes occurring in the ruptured follicle in the lower Vertebrata, and

also in the Monotremes. According to Sandes, [Footnote: 'The Corpus Luteum

of Dasyurus,' _Proc. Lin. Soc._, New South Wales, 1903.] in the latter

there is a pronounced hypertrophy of the follicular epithelium after

ovulation, but no ingrowth of connective tissue or blood-vessels from the

follicular wall. Marshall himself examined sections of the corpus luteum

of _Ornithorhynchus_ and saw much hypertrophied and apparently fully

developed luteal cells, but no trace of any ingrowth from the wall of the

follicle. This fact would appear to be quite inconsistent with the theory

above proposed, but it must be remembered that the ovum of Monotremes is

known to remain for a short period in the oviduct, or in other words to

pass through it very slowly, and to absorb fluid from its walls, as shown

by the considerable increase in size which the ovarian ovum undergoes

before it is laid. It would be interesting to know how long the

rudimentary corpus luteum persists in _Ornithorhynchus_: the period,

according to my views, should be very short. It is remarkable that in the

results quoted by Marshall a well-developed corpus luteum was found and

exclusively found in the lower Vertebrates which are viviparous. For

example, among fishes in the Elasmobranchs _Myliobatis_ and _Spinax_; in

Teleosteans, in _Zoarces_; in Reptiles, in _Anguis_ and _Seps_. Bühler on

the other hand, confirmed my own negative result with regard to oviparous

Teleosteans, and also found no hypertrophy of the follicle in Cyclostomes

which are also oviparous. In the viviparous forms mentioned there is yolk

in the ovum which is retained in oviduct or ovary, but additional

nutriment is also absorbed from the uterine or ovarian walls. In these

cases there is no placenta and generally no adhesion of ovum or embryo to

walls of oviduct or ovary. These facts alone would be sufficient to

disprove the theory that the corpora lutea are organs producing a

secretion whose function is to cause the attachment of the embryo to the

uterine mucosa. It is also, in my opinion, unreasonable to suppose that

the rudimentary corpora lutea of lower viviparous Vertebrates arose as a

mutation the result of which was to cause internal development of the

ovum. Habits might easily bring about retention of the fertilised ova for

gradually increasing periods, [Footnote: According to Geddes and Thomson

(_Evolution of Sex_, 1889), the common grass-snake has been induced under

artificial conditions to bring forth its young alive.] and the correlation

between the retained developing ova and the hypertrophy of the ruptured

follicles is comprehensible on my theory of the influence of substances

absorbed by the walls of oviduct or ovary from the developing ovum.

The case of _Dasyurus_, however, seems inconsistent with this argument,

for, as previously mentioned, Sandes found that in this Marsupial the

corpora lutea persisted during the greater part of the period of

lactation, which continues for four months after parturition. During the

whole of this time there are no embryos in the uteri, and therefore it

might be urged absorption of hormones from the embryos cannot be the cause

of the persistence of corpora lutea in pregnancy. But it seems to me that

a complete answer to this objection is supplied by the peculiar relations

of the embryos to the pouch in _Dasyurus_ and other Marsupials. The skin

of the pouch while the embryos are in it is very soft, congested, and

glandular; at the same time the embryos when transferred to the pouch at

parturition are very small, immature, and have a soft delicate skin. The

relation of embryos to pouch in _Dasyurus_, therefore, is closely similar

to that of embryos to uterus after the first few days of pregnancy in the

Eutheria. It is true there is no placenta, but the mouths of the embryos

are in very close contact with the teats, and both the skin of the embryos

and that of the pouch are soft and moist. If any special substances are

given off by the embryos in the uterus in ordinary gestation, the same

substances would continue to be given off by the embryos in the marsupial

pouch, and these must be absorbed by the skin of the pouch. In this way it

seems to me we have a logical explanation of the fact that the corpora

lutea in the Marsupial are not absorbed at parturition as in Eutheria. As

Sandes says the 'greater part of the period of lactation,' it would appear

that absorption of the corpora lutea takes place when the young _Dasyurus_

have grown to some size, become covered with hair, and are able to leave

the teats or even the pouch at will. Under these conditions it is obvious

that diffusion of chemical substances from the young through the walls of

the pouch would come to an end. It would be interesting in this connexion

to know more of the relation of egg and embryo to the pouch and to the

corpora lutea in _Echidna_. In _Ornithorhynchus_ the eggs are hatched in a

nest and there is no pouch.

On this view that the corpora lutea are the result, not the cause, of

intra-uterine gestation, it would no longer be possible to maintain the

theory that the corpus luteum in the human species is the cause by its

internal secretion of the phenomenon of menstruation. This was the theory

of Born and Fränkel. [Footnote: See Biedl, _Internal Secretory Organs_

(Eng. trans.), 1912, p. 404.] Biedl's conclusion is that the periodic

development and disintegration of the uterine mucous membrane in the

menstrual cycle is due to the hormone of the interstitial cells of the

ovary. Leopold and Ravana found that ovulation as a rule coincides with

menstruation, but may take place at any time. Here, again, the problem

must be considered from the point of view of evolution. It can scarcely be

doubted that the thickening and growth of the mucous membrane in the

menstrual cycle is of the same nature as that which takes place in

pregnancy. When the ovum or ova are not fertilised the development comes

to an end after a certain time, differing in different species of Mammals,

and the membrane sloughs, returns to its original, state, and then begins

the same process of development again.

Menstruation, then, must be interpreted as an abortive parturition, both

in woman and lower Mammals, though in the latter it is not usually

accompanied by hemorrhage, and is called pro-oestrus. The question then to

be considered is, what determines parturition and menstruation? The

presence of the fertilised ovum must have been the original cause of the

hypertrophy of the uterine mucous membrane, and in its congenital or

hereditary development the chemical substances diffusing from the ova in

the uterus or even in the Fallopian tube may well be the stimulus starting

the hypertrophy. But what determines the end of the pregnancy? Is it

merely the increasing distension of the uterus by the developing foetus?

This could scarcely be the case in the Marsupials in which the foetus when

born is quite minute. Nor can we attribute parturition to renewed

ovulation, for this occurs in _Dasyurus_ only once a year. All we can

suggest at present is that a certain periodic development takes place by

heredity in presence of the hormones exuded by the fertilised ovum and the

embryo developed from it. When the ovum or ova, not being fertilised, die

the period of development is (usually) shortened and pro-oestrus or

menstruation occurs. In the dog, however, the period of the oestrus cycle

is about the as that of gestation--namely, six months.

The so-called descent of the testicles occurs exclusively in Mammals, in

which with a few important exceptions it is universal. This is a very

remarkable case of the change of position of an organ in the course of

development. The original position of the testis on either side is quite

similar to that of the same organ in birds or reptiles. The genital ridge

runs along the inner edge of the mesonephros, with which the testicular

tubules become connected. The testis, with the mesonephros, forming the

epididymis, closely attached to it, projects into the coelom, and without

losing its connexion with the peritoneum changes its position gradually

during development, passing backwards and downwards until it comes to lie

over the wall of the abdomen just in front of the pubic symphysis of the

pelvic girdle. There the abdominal wall on either side of the middle line

becomes thin and distended to form a pouch, the scrotal sac, into which

the testis passes, still remaining attached to the peritoneum which lines

the pouch, while the distal end of the vas deferens retains its original

connexion with the urethra. The movement of the testis can thus be

accurately described as a transposition or dislocation.

Various causes have been suggested for the formation of the scrotum, but

no one has ever been able to suggest a use for it. It has always been

quite impossible to bring it within the scope of the theory of natural

selection. The evolution of it can only be explained either on the theory

of mutation or some Lamarckian hypothesis. The process of dislocation of

the testis does not conform to the conception of mutation, nor agree with

other cases of that phenomenon. A mutation is a change of structure

affecting more or less the whole soma, but showing itself especially in

some particular organ or structure. But I know of no mutation occurring

under observation which consisted, not in a change of structure or

function, but merely in a change of position of an organ from one part of

the body to another, and moreover a change which takes place by a

continuous process in the course of development. If the testes were

developed from the beginning in a different part of the abdomen, there

might be some reason in calling the change a mutation. Moreover, if it is

a mutation, why has it never occurred in any other class of Vertebrates

except Mammals?

In 1903 Dr. W. Woodland published [Footnote: _Proc. Zool. Soc._, 1903,

Part 1.] a Lamarckian theory of this mammalian feature, the probability of

which it seems to me has been increased rather than decreased by the

progress of research concerning heredity and evolution since that date.

Dr. Woodland correlated the dislocation of the testes with the special

mechanical features of the mode of locomotion in Mammalia. His words are:

'The theory here advocated is to the effect that the descent of the testes

in the Mammalia has been produced by the action of mechanical strains

causing rupture of the mesorchial attachments, such strains being due to

the inertia of the organs reacting to the impulsiveness involved in the

activity of the animals composing the group.' The 'impulsiveness' is the

galloping or leaping movement which is characteristic of most Mammals when

moving at their utmost speed, as seen, for example, in horses, deer,

antelopes, dogs, wolves, and other Ungulata and Carnivora. It is obvious

that when the body is descending to the ground after being hurled upwards

and forwards, the abdominal organs have acquired a rapid movement

downwards and forwards; when the body reaches the ground its movement is

stopped suddenly, while the abdominal organs continue to move. The testes

therefore are violently jerked downwards away from their attachments and

at the same time forward. The check to the forward movement, however, is

momentary, while the body is immediately thrown again upwards and

forwards, which by the law of inertia means that the testes are thrown

still more downwards and backwards. There is no reason to suppose, as Dr.

Woodland suggests, that any rupture of the mesorchium was the usual result

of these strains, but a constant pull or tension was caused in the

direction in which the testes actually move during development. On this

theory we have to consider (1) how such strains could cause a shifting of

the peritoneal attachment, (2) why the testes should be supposed to be

particularly affected more than other abdominal organs. The answer to the

first question is that the strains would cause a growth of the connecting

membrane (mesorchium) at the posterior end, accompanied by an absorption

of it at the anterior end. The answer to the second question is that the

testes are at once the most compact and heaviest organs in the abdomen,

and at the same time the most loosely attached. The latter statement does

not apply to the mesonephros or epididymis which has moved with the

testis, but the latter cannot function without the former, and it may be

supposed that the close attachment of the epididymis to the testis had

come about in the early Mammalia before the change of position was

evolved.

It is evident that the violent shocks of the galloping or leaping movement

do not occur in Birds, Reptiles, or Amphibia. Ostriches run very fast and

do not fly, but their progression is a stride with each foot alternately,

not a gallop. The Anura among the Amphibia are saltatory, but their leaps

are usually single, or repeated only a few times, not sustained gallops.

The exceptions among the Mammalia still more tend to prove the close

correspondence between the 'impulsive' mode of progression and the

dislocation of the male gonads. In the Monotremata there is no scrotum,

the testes are in a position similar to that which obtains in Reptiles,

and they are the only Mammals in which these organs are anterior to the

kidneys. In locomotion they are sluggish, there is no running or galloping

among them. _Ornithorhynchus_ is aquatic in its habits, and _Echidna_ is

nocturnal and moves very slowly. In Marsupials the scrotum is in front of

the penis, but really in the same position as in other Mammals--that is,

in front of the ventral part of the pelvic girdle. It is the penis which

is different, as the skin around the organ has not united in a ventral

suture below it, while the organ itself has not grown forward adnate to

the abdominal skin as in most other Mammals. The scrotum is always

anterior to the origin of the penis, although in the Eutheria apparently

behind that organ. The larger Marsupials like the kangaroos are eminently

saltatory, and the others are active in locomotion. The aquatic Mammals

Sirenia and Cetacea have no scrotum, the testes being abdominal. It is

unnecessary to inquire whether this is the original position, or whether

they are descended from ancestors which had a scrotum: in either case the

position of the testes corresponds to the absence of what Dr. Woodland

calls impulsiveness in progression. The Fissipedia offer an instructive

example, for while the Otariidae have the hind feet turned forward and can

move on land somewhat like ordinary Mammals, the Phocidae cannot move

their hind legs independently or turn them forward, and can only drag

themselves about on land for short distances. In the former the testes are

situated in a well-defined scrotum, in the latter these organs are

abdominal. The Phocidae are probably descended from Mammals of the

terrestrial type with a scrotum, which has disappeared in the course of

evolution. Perhaps the most curious exception is that of the elephants, in

which the testes are abdominal. Here, in consequence of their structure

and massive shape, locomotion in usually a walk, and though they run

occasionally the gait is a trot, not a sustained gallop, and leaping is

out of the question. Sloths which hang from branches upside down have

abdominal testes, but even here they are in a posterior position, between,

the rectum and the bladder, so there has apparently been a degree of

dislocation, probably inherited from ancestors with more terrestrial

habits.

The fact that the ovaries do not occupy normally a position similar to

that of the testes is in accordance with the theory, for they are very

much smaller than the testes; and yet they have undergone some change of

position, for they are posterior to the kidneys.

The facts agree with the hormone theory, for it is to be noted that

although the development of the scrotum is confined to the males, the

'descent' or dislocation takes place in the foetus, and not at the period

of puberty. This is in accordance with the fact that the mechanical

conditions to which the change is attributed are not related to sexual

habits, but to the general habits of life which begin soon after birth.

The development, therefore, may be considered to be related to the

presence of a hormone derived from the normal testis, but not to a special

quantity or quality of hormone associated with maturity or the functional

activity of the organ. In Rodents, however, there is a difference in the

organs, not only at maturity, but in every rutting season, at any rate in

Muridae such as rats and others. In the rutting season the testes become

much larger and descend into the scrotal sacs, at other times of the year

being apparently more or less abdominal. In rabbits and hares, which have

a much more impulsive progression, the organs seem to be always in the

scrotal sacs.

It might be thought that in this case, although the hormone theory of

heredity might be applied, there was no reason to suppose that a hormone

derived from the testis in the individual development was necessary in

order that the hereditary change should take place. If the individual was

male and therefore had a testis, this organ would by heredity go through

the process of dislocation. But there is the curious fact that when the

descent is not normal and complete, in what is called cryptorchidism, the

organs are always sterile. The retention of the testes within the abdomen

may be regarded as a case of arrested development, like many other

abnormalities, but this does not explain why the retained testes should

always be sterile, without spermatogenesis. If the inherited or congenital

process of dislocation requires the presence of hormones produced by a

normal testis, then we can understand why a defective testis does not

descend completely, because it does not produce the hormone which is

necessary to stimulate the hereditary mechanism to complete dislocation.

It is often stated that in cryptorchidic individuals the sexual instincts

and somatic sexual characters are well developed, which would appear

contradictory to the above explanation, but according to Ancel and Bouin

such individuals in the case of the pig show considerable differences in

the secondary signs of sex and in the external genital organs, presenting

variations which lie between the normal and the castrated animal.

We have here, then, in the position of the testes in Mammalia a condition

which is not in the slightest degree 'adaptive' in the ordinary sense--

that is, fulfilling any special function or utility. The condition must be

regarded as distinctly disadvantageous, since the organs are more exposed

to injury, and the abdominal wall is weakened, as we know from the risk of

scrotal hernia in man. But from the Lamarckian point of view the facts

support the conclusion that the condition is the effect of certain

mechanical strains, and is of somatic origin, while the correlations here

reviewed are entirely unexplained by any theory of mutation or blastogenic

origin.

OPPOSING EVIDENCE

We have now to review certain cases which seem to support conclusions

contrary to those which we have maintained in the preceding pages, and to

consider the evidence which has been published in support of other

theories. It must be admitted that the occurrence of male secondary

characteristics on one side of the body, and female on the other, is in

consistent with the view that the development of such characters is due to

the stimulus of a hormone, since the idea of a hormone means something

which diffuses by way of the blood-vessels, lymph-vessels, and interstices

of the tissues, throughout the body, and the hormone theory of secondary

sexual characters assumes that these characters are potentially present by

heredity in both sexes. The occurrence of male somatic characters on one

side or in some part of the body and female on the other, usually

associated with the corresponding gonads, has been termed

gynandromorphism, and has long been known in insects. Cases of this

condition have been observed, though much more rarely, in Vertebrates. I

am not aware of any authentic instances in Mammals, and the supposition

that in stags reduction or abnormality of one antler may be the result of

removal or injury to the testis of one side, or the opposite, have been

completely disproved by experiments in which unilateral castration has

been carried out without any effect on the antlers at all. In birds,

however, a few cases have been recorded by competent observers with a

definiteness of detail which leaves no possibility of doubt. One of the

more recent of these is that of a pheasant of the white-ringed Formosan

variety, _P. torquatus_, of the Chinese pheasant. [Footnote: C. J. Bond,

'Unilateral Development of Secondary Male Characters in a Pheasant,'

_Journ. of Genetics_, vol. iii., 1914.] On the left side this bird shows

the plumage, colour, and the spur of the male; on the right leg there is

no spur except the small rudiment normally occurring in the hen. The

difference in plumage between the two sides, however, is not complete. The

white collar is strictly limited to the left side, but the iridescent blue

green of head and neck is present on both sides, though more marked on the

left. Only a few male feathers appear in the wing coverts of the left

side. The breast feathers are rufous, especially on the left side. The

tail coverts show marked male characters, more especially on the left

side. In the tail, however, the barred character of the male is not

present on one side, absent on the other, but in most of the feathers is

confined to one, the _outer_ side of each feather. With regard to the

gonads, in this bird a single organ was found on the left side, _i.e._ in

the position of the ovary in normal females, and there was no trace of a

gonad on the right side. The organ present was small, 3/4 inch long by 1/2

inch broad, and microscopic sections showed in one part actively growing

areas of tubular gland structure in some of which bodies like spermatozoa

could be detected, while in another were fibrous tissue with degenerating

cysts. The latter appear to have been degenerating egg follicles. The

author concludes that the organ was originally a functional ovary, and

that the ovarian portion had atrophied while a male portion had become

functionally active.

Another case in birds was described by Poll [Footnote: _B.B. Ges. Naturf.

Freunde_, Berlin, 1909.] and is mentioned by Doncaster. [Footnote:

_Determination of Sex_, Cambridge, 1914.] It is that of a Bullfinch

which had the male and female plumage sharply separated on the two sides

of the body. The right side of the ventral surface was red like a normal

male, the left side grey like a normal female. In this case there was a

testis on the right side, on the left an ovary as in normal females.

A third case in birds, somewhat different from the two first mentioned, is

that of a domestic fowl described by Shattock and Seligmann. [Footnote:

_Trans. Pathol. Soc._ (London), vol. 57, Part i., 1906.] It was a bird of

the Leghorn breed, two years old, and had the fully developed comb and

wattles of the cock. Each leg bore a thick blunt spur, nearly an inch in

length, but in the Leghorn breed spurs are by no means uncommon in hens of

mature age, before they have ceased to lay eggs. In plumage the characters

were mainly female. The colour being white could not show sexual

differences, the neck hackles were but moderately developed, saddle

hackles practically absent, the tail resembled that of the hen. There was

a fully developed oviduct on the left side, on the right another less than

half the full length. There was also a vas deferens on each side. There

was a gonad on each side, that of the right about one-fourth the size of

that on the left. In microscopic structure the right gonad resembled a

testis consisting entirely of tubuli lined by an epithelium consisting of

a single layer of cells. In one part of this organ the tubules were larger

than elsewhere, and one of them exhibited spermatogenesis in progress. The

left and larger gonad had a quite similar structure, but at its lower end

were found two ova enclosed within a follicular epithelium.

With regard to the last case it is to be remarked that though the gonad on

the right side was entirely male, there was no unilateral development of

male characters. With regard to the other two cases it must be pointed out

(1) that the difference between the two somatic sex-characters on the two

sides is chiefly a difference of colour, except the difference in the

spurs in Bond's pheasant; (2) that the evidence already cited shows that

in fowls castration does not prevent the development of the colour and

form of the male plumage, nor of the spurs: that in drakes, although

castration does not seem to have been carried out on young specimens

before the male plumage was developed, when performed on the mature bird

it prevents the eclipse, and does not cause the male to resemble the hen.

Castration, then, tends to prove that in Birds the development of the male

characters is not so closely dependent on the stimulation of testicular

hormone as in Mammals. The characters must therefore be developed by

heredity in the soma, which implies that the soma must itself be

differentiated in the two sexes. The development must therefore be more

in the nature of gametic coupling. It does not follow that the primary

sex-character or the somatic characters are exclusive in either sex.

We may suppose that the zygote contains both sexes, one or other of which

is dominant, and that dominance of one primary sex involves dominance of

the corresponding sexual characters. This does not, however, agree with

the result of removal of the ovaries in ducks, for this causes the

characters of the male to appear, so that the dominance of the female is

not a permanent condition of the soma but is dependent on the ovarian

hormone.

In the hermaphrodite individuals mentioned above the difference of

dominance is on two sides of the body instead of two different

individuals. It may also be remarked here that while it is very difficult

to believe that spurs were not due in evolution to the mechanical

stimulation of striking with the legs in combat, and while specially

enlarged feathers are erected in display, we cannot at present attribute

the varied and brilliant _colour_ of male birds to the direct influence of

external stimuli.

In Lepidoptera among insects the evidence concerning castration tends to

prove that hormones from the gonads play no part at all in the development

of somatic sexual characters. Kellog, an American zoologist, in 1905

[Footnote: _Journ. Exper. Zool._ (Baltimore), vol. i., 1905.] described

experiments in which he destroyed by means of a hot needle the gonads in

silkworm caterpillars (_Bombyx mori_), and found no difference in the

sexual characters of the moths reared from such caterpillars. Oudemans had

previously obtained the same result in the Gipsy Moth, _Limantria dispar_.

Meisenheimer [Footnote: _Experimentelle Studien zur Soma- und

Geschlechtedifferenzierung_. Jena, 1909.] made more extensive

experiments on castration of caterpillars in the last-mentioned species,

in which the male is dark in colour and has much-feathered antennae, while

the female is very pale and has antennae only slightly feathered. In the

moths developed from the castrated larvae there was no alteration in the

male characters, and in the females the only difference was that some of

them were slightly darker than the normal. Meisenheimer and Kopee after

him claim to have grafted ovaries into males and testes into females, with

the result that the transplanted organs remained alive and grew, and in

some cases at least became connected with the genital ducts. Even in these

cases the moth when developed showed the original characters of the sex to

which belonged the caterpillar from which it came, although it was

carrying a gonad of the opposite sex. It will be seen that these results

are the direct opposite of those obtained by Steinach on Mammals. We have

no evidence that the darker colour of the normal male in this case is

adaptive, or due to external stimuli, but the feathering of the antennae

is generally believed to constitute a greater development of the olfactory

sense organs, and is therefore adaptive, enabling the male to find the

female. This is therefore the kind of organ which would be expected to be

affected by hormones from the generative organs. It is stated that the

sexual instincts were also unaltered, a male containing ovaries instead of

testes readily copulating with a normal female.

These results, almost incredible as they appear, are in harmony with the

relatively frequent occurrence of gynandromorphism in insects.[Footnote:

See Doncaster, _Determination of Sex_ (Camb. Univ. Press, 1914), chap.

ix.] One of the most remarkable cases of this is that of an ant

(_Myrmica scabrinodis_) the left half of which is male, the right half not

merely female, but worker--that is, sterile female, without wing. Cases in

Lepidoptera, _e.g. Amphidasys betularia_, have frequently been recorded.

Presumably not only the antennae and markings, but also the genital

appendages and the gonads themselves, are male and female on the two

sides. On the view that both sexes and the somatic sex-characters of both

sexes are present in each zygote, and that the actual sex is due to

dominance, we must conclude that the male primary and secondary characters

are dominant on one side, and the female on the other, and it is evident

that hormones diffusing throughout the body cannot determine the

development of somatic sexual characters here. Various attempts have been

made to explain gynandromorphism in insects in accordance with the

chromosome theory of sex-determination. These are discussed by Doncaster

in the volume already cited, but from the point of view of the present

work the important question is that concerning the somatic sex-characters.

According to Doncaster it has been found that in some Lepidoptera the

different sex-chromosomes occur in the female, not in the male as in other

insects. Half the eggs, therefore, contain an X chromosome, and half a Y,

while all the sperms contain an X chromosome. Doncaster has seen in

_Abraxas grossulariata_ ova with two nuclei both undergoing maturation.

If one of these in reduction expelled a Y chromosome, the other an X,

then one would retain an X and the other a Y. Each was fertilised by a

sperm, one becoming therefore XX or male and the other XY or female. It

may be supposed that as there was only the cytoplasm of one ovum, each

nucleus would determine the characters of half the individual developed.

The question remains, therefore, where are the factors of the somatic

sex-characters? One suggestion which might be made is that the female

characters are present in the _Y_, in this case female producing

chromosome, or, if the female characters are merely negative, that the

male characters are in the _X_ chromosome, but only show themselves in the

homozygous condition, thus:--

FEMALE x MALE

XY XX

| \/ |

| /\ |

XX YX

MALE FEMALE

The male characters in the male, _XX_, would appear because present in two

chromosomes, but would be recessive in the female because present only in

one chromosome. The validity of this scheme, however, is disproved by the

fact that males can transmit the female characters of their race, as in

the case mentioned by Doncaster where a male _Nyssia zonaria_ when crossed

transmits the wingless character of its own female.

Another, perhaps better, suggestion is that the somatic characters of both

sexes are present in each. Then as each somatic cell is descended without

segregation from the fertilised ovum, we may suppose that the presence of

the sex-chromosomes in the somatic cells themselves in some way determines

whether male or female characters shall develop, without the aid of any

hormones from the gonads. This theory would be quite compatible with

the belief that adaptive somatic sex-characters may be due to external

stimulation, for supposing that the hypertrophy or modification is

conveyed to the determinants in the gametocytes, and was confined to

one sex, _e.g._ the male, then these determinants would be modified in

association with the sex-chromosomes of that sex, and thus though

after reduction and fertilisation they would be present in the female

zygote also, they would not develop in that sex. Thus supposing _M_ to

represent a modification acquired in the male and _m_ the absence of

the modification, such as the feathered antenna of a moth, and the

sex-chromosomes to be _X_ and _Y_, then we should have in the

gametocytes--

Male Female

_MM mm_

_XX XY_

Gametes _MX, MX: mXmY_

Zygotes _MmXX male, MmXY female_,

and the character _M_ would only appear in the male because it only

develops in association with _XX_ in the somatic cells descended from the

male zygote. This would be the result in the first generation in which a

somatic modification affected the factors in the chromosomes. In the next

generation _m_ in the male would be affected, and the male for the sake of

simplicity might be supposed to become _MMXX_. When the female gametes

segregated, some would always be _mY_, and some zygotes therefore _MXmY_.

Others might be _MMXY_. On this theory, therefore, there would always be

some females heterozygous for the male character.

Geoffrey Smith, one of the many promising young scientific investigators

whose careers were cut short in the War, maintained views concerning

somatic sex-characters different from that which explains their

development as due to a hormone from the testis or ovary. Nussbaum in 1905

[Footnote: 'Ergebuisse der Anat. und Entwicklungsgesch.,' Bd. xv.;

_Pflügers Archiv_, Bd. cxxvi, 1909.] had recorded experiments on _Rana

fusca_ (which is identical with the British species commonly called _R.

temporaria_) which appeared to prove that in the male frog after

castration the annual development of the thumb-pad and the muscles of the

fore-leg does not take place, and if these organs have begun to enlarge

before castration they atrophy again. When pieces of testis were

introduced into the dorsal lymph-sac of a castrated frog the thumb-pads

and muscles developed as in a normal frog. Geoffrey Smith and Edgar

Schuster [Footnote: _Quart Journ. Mic. Sci_., lvii, 1911-12.] investigated

the subject again with results contrary to those of Nussbaum.

Smith and Schuster begin by describing the normal cycle of changes in the

testes on the one hand and the thumb-pad on the other. After the discharge

of the spermatozoa in March or April the testes are at their smallest

size. From this time onwards till August they steadily increase in size,

attaining their maximum at the beginning of September. From then till the

breeding season no increase in size or alteration of cellular structure

occurs, the testes apparently remaining in a state of complete inactivity

during this period. With regard to internal development, after the

discharge of spermatozoa in the breeding season the spermatogonia divide

and proliferate, forming groups of cells known as spermatocysts. In June

and July spermatogenesis is active, and from August to October the

formation of ripe spermatozoa is completed.

The corresponding changes in the thumb-pads are as follows. Immediately

after the breeding season the horny epidermis of the pad with its deeply

pigmented papillae is cast off, and the thumb remains comparatively smooth

from April or May until August or September. When the large papillae are

shed, smaller papillae remain beneath, and are gradually obliterated by

the epidermis growing up between them. The epidermis is therefore growing

while the spermatogenesis is taking place. In August and September the

epidermic papillae begin to be obvious, and from this time till February a

continuous increase in the papillae and their pigmentation occur. Geoffrey

Smith argues that the development of this somatic character occurs while

the testes are inactive and unchanged. Considering that the testes

throughout the winter months are crammed with spermatozoa, which must

require some nourishment, and which may be giving off a hormone all the

time, the argument has very little weight. Smith and Schuster found that

ovariotomy, with or without subsequent implantation of testes or injection

of testis extract, had no effect in causing the thumb of the female to

assume any male characters.

Castration during the breeding season causes the external pigmented

layer with its papillae to be cast off very soon--that is to say, it

has the same effect as the normal discharge of the spermatozoa. Smith

and Schuster found that castration at other seasons caused the pad to

remain in the condition in which it was at the time, that there was no

reduction or absorption as Nussbaum and Meisenheimer found, and that

allo-transplantation of testes--that is, the introduction of testes from

other frogs either into the dorsal lymph-sacs or into the abdominal

cavity--or the injection of testis extract, had no effect in causing

growth or development of the thumb-pad.

There seems to be one defect in the papers of both Nussbaum and Smith and

Schuster--namely, that neither of them mentions or apparently appreciates

the fact that the thumb-pads, apart from the dermal glands, consist of

horny epidermis developed from the living epidermis beneath. The horny

layer is not shown clearly in the figures of Smith and Schuster. It seems

impossible that the horny layer or its papillae could atrophy in

consequence of castration, or be absorbed. The horny part of the frog's

thumb-pad is comparable with the horny sheath of the horns in the

mammalian Prong-buck (_Antilocapra_) which are shed after the breeding

season and annually redeveloped. Meisenheimer claims that he produced

development of papillae on the thumb-pad, not only by implantation of

pieces of testis, but also by implantation of pieces of ovary. This seems

so very improbable that it suggests a doubt whether the same investigator

was not mistaken with regard to the results of his experiments in

transplanting gonads in Moths.

Smith and Schuster conclude that the normal development of the thumb-pad

depends on the presence of normal testes, but that there is no sufficient

evidence that the effect is due to a hormone derived from the testis. It

is equally probable, according to Smith, that the testicular cells take up

some substance or substances from the blood, thus altering the composition

of the latter and perhaps stimulating the production of these substances

in some other organ of the body. These substances may be provisionally

called sexual formative substances. Smith's theory therefore is that the

action of the testes in metabolism is rather to take something from the

blood than to add something to it, and that it is this subtractive effect

which influences the development of somatic sexual organs.

Geoffrey Smith in fact, in the paper above considered, attempts to apply

to the frog the views he put forward [Footnote: _Fauna und Flora des

Golfes van Neapel_, 29 Monographie Rhizocephala.] in relation to the

effect of the parasite _Sacculina_ on the sexual organs of crabs. The

species in which he made the most complete investigation of the influence

of the parasite was _Inachus scorpio_ (or _dorsettensis_). Figures showing

the changes in the abdomen produced by the presence of _Sacculina_ are

given in Doncaster's _Determination of Sex_, Pl. xv. _Sacculina_ is one of

the Cirripedia, and therefore allied to the Barnacles. It penetrates into

the crab in its larval stage, and passes entirely into the crab's body,

where it develops a system of branching root-like processes. When mature

the body of the _Sacculina_ containing its generative organs forms a

projection at the base of the abdomen of the crab on its ventral surface,

and after this is formed the crab does not moult. Crabs so affected do not

show the usual somatic sexual characters, and at one time it was supposed

that only females were attacked. It is now known that both sexes of the

host may be infected by the parasite, but the presence of the latter

causes suppression of the somatic sex-differences. The entry of the

parasite is effected when the crab is young and small, before the somatic

sex-characters are fully developed. The gonads are not actually

penetrated, at least in some cases, by the fibrous processes of the

parasite, but nevertheless they are atrophied and almost disappear. In

_Inachus_ the abdomen of the normal male is very narrow and has no

appendages except two pairs of copulatory styles. The abdomen of the

female is very broad, and has four pairs of biramous appendages covered

with hairs, the normal function of which is to carry the eggs. The effect

of the parasite in the male is that the abdomen is broader, the copulatory

styles reduced, and biramous hairy appendages are developed similar to

those of the female, but smaller. In the female the abdomen remains broad,

but the appendages are much smaller than in the normal female, about equal

in size to those of the 'sacculinised' male. Smith interpreted the

alteration in the male as a development of female secondary characters,

but it is obvious from the condition in Macrura or tailed Decapods, like

the lobster or crayfish, that the abdomen or tail of the male originally

carried appendages similar to those of the female, and that the male

character is a loss of these appendages. The absence of the male character

therefore necessarily involves a development of these appendages, and

there is not much more reason for saying that the male under the influence

of the parasite develops female characters, than for saying that the male

character is absent. There is no evidence in the facts concerning

parasitic castration for Geoffrey Smith's conclusion that the female

characters are latent in the male, but the male characters not latent in

the female: both return to a condition in which they resemble each other,

and the primitive form from which they were differentiated.

By his studies of parasitic castration Geoffrey Smith was led to formulate

a theory for the explanation of somatic sex-characters different from that

of hormones. He found that in the normal female crab the blood contained

fatty substances which were absorbed by the ovaries for the production of

the yolk of the ova. When _Sacculina_ is present these substances are

absorbed by the parasite; the ovary is deprived of them, and therefore

atrophies. In the male the parasite requires similar substances, and its

demand on the blood of the host stimulates the secretion of such

substances, so that the whole metabolism is altered and assimilated to

that of the female. It is this physiological change which causes the

development of female secondary characters. He describes this change as

the production of a hermaphrodite sexual formative substance, on the

ground that in at least one case eggs were found in the testis of a male

_Inachus_ which had been the host of a _Sacculina_, but had recovered. It

must however be noted that the _Sacculina_ itself is hermaphrodite, with

ovaries much larger than the testes. It is possible that while the

parasite prevents the development of testis or ovary in the host, it gives

up to the body of the host a hormone from its own ovaries which tends to

develop the female secondary characters: for the parasite is itself a

Crustacean, and therefore the hormone from its ovaries would not be of too

different a nature to act upon the tissues of the host.

The observation of Geoffrey Smith that eggs may occur in the testis of a

crab after recovery from the parasite appears of more importance than his

peculiar theoretical suggestions, for it tends to show that sex is not

always unalterably fixed at fertilisation. In this case the influence of a

parasite predominantly female would seem to be the real cause of the

development of eggs in the testis of the host. Geoffrey Smith does not

discuss the origin of the somatic sexual characters in evolution, or

attempt to show how his theories of sexual formative substance, and of the

influence of the gonads by subtraction rather than addition, would bear

upon the problem.

CHAPTER VI

Origin Of Non-Sexual Characters: The Phenomena Of Mutation

According to the theory here advocated, modifications produced by external

stimuli in the soma will also be inherited in some slight degree in each

generation when they have no relation to sex or reproduction. In this case

the habits and the stimuli which they involve will be common to both

sexes, and the hormones given off by the hypertrophied tissues will act

upon the corresponding determinants in the gametocytes. The modifications

thus produced will therefore be related to habits, and the theory will

include all adaptations of structure to function, but other characters may

also be included which are the result of stimuli and yet have no function

or utility.

The majority of evolutionists in recent years have taught that influences

exerted through the soma have no effect on the determinants in the

chromosomes of the gametes, that all hereditary variations are gametogenic

and none somatogenic. Mendelians believe that evolution has been due to

the appearance of characters or factors of the same kind as those which

distinguish varieties in cultivated organisms, and which are the subject

of their experiments, but they have found a difficulty, as already

mentioned in Chapter II, in forming any idea of the origin of a new

dominant character. A recessive character is the absence of some positive

character, and if in the cell-divisions of gametogenesis the factor for

the positive character passes wholly into one cell, the other will be

without it, will not 'carry' that factor. If such a gamete is fertilised

by a normal gamete the organism developed from the zygote will be

heterozygous, and segregation will take place in its gametes between the

chromosome carrying the factor and the other without it, so that there

will now be many gametes destitute of the factor in question. When two

such gametes unite in fertilisation the resulting organism will be a

homozygous recessive, and the corresponding character will be absent. In

this way we can conceive the origin of albino individuals from a coloured

race, supposing the colour was due to a single factor.

In Bateson's opinion the origin of a new dominant is a much more difficult

problem. In 1913 he discussed the question in his Silliman Lectures.

[Footnote: _Problems of Genetics_, Oxford Univ. Press, 1913.] He considers

the difficulty is equally hopeless whether we imagine the dominants to be

due to some change internal to the organism or to the assumption of

something from without. Accounts of the origin of new dominants under

observation in plants usually prove to be open to the suspicion that the

plant was introduced by some accident, or that it arose from a previous

cross, or that it was due to the meeting of complementary factors. In

medical literature, however, there are numerous records of the spontaneous

origin of various abnormalities which behave as dominants, such as

brachydactyly, and Bateson considers the authenticity of some of these to

be beyond doubt. He concludes that it is impossible in the present state

of knowledge to offer any explanation of the origin of dominant

characters. In a note, however, he suggests the possibility that there are

no such things as new dominants. Factors have been discovered which simply

inhibit or prevent the development of other characters. For example, the

white of the plumage in the White Leghorn fowl is due to an inhibiting

factor which prevents the development of the colour factor which is also

present. Withdraw the dominant inhibiting factor, and the colour shows

itself. This is shown by crossing the dominant white with a recessive

white, when some birds of the F(2) generation are coloured.[Footnote:

Bateson, _Principles of Heredity_, p. 104.] Similarly, brachydactyly in

man may be due to the loss of an inhibiting factor which prevents it

appearing in normal persons. It is evident, however, that it is difficult

to apply this suggestion to all cases. For example, the White Leghorn fowl

must have descended from a coloured form, probably from the wild species

_Gallus bankiva_. If Bateson's suggestion were valid we should have to

suppose that the loss of the factor for colour caused the dominant white

to appear, and then when this is withdrawn colour appears again, so that

the colour factors and the inhibiting factors must lie over one another in

a kind of stratified alternation. And then how should we account for the

recessive white?

In his Presidential Address to the meeting of the British Association in

Australia, 1914, Bateson explains his suggestion somewhat more fully with

a command of language which is scarcely less remarkable than the subject

matter. The more true-breeding forms are studied the more difficult it is

to understand how they can vary, how a variation can arise. When two forms

of _Antirrhinum_ are crossed there is in the second generation such a

profusion of different combinations of the factors in the two

grandparents, that Lotsy has suggested that all variations may be due to

crossing. Bateson does not agree with this. He believes that genetic

factors are not permanent and indestructible, but may undergo quantitative

disintegration or fractionation, producing subtraction or reduction

stages, as in the Picotee Sweet Pea, or the Dutch Rabbit. Also variation

may take place by loss of factors as in the origin of the white Sweet Pea

from the coloured. But regarding a factor as something which, although it

may be divided, neither grows nor dwindles, neither develops nor decays,

the Mendelian cannot conceive its beginning any more than we can conceive

the creation of something out of nothing. Bateson asks us to consider

therefore whether all the divers types of life may not have been produced

by the gradual unpacking of an original complexity in the primordial,

probably unicellular forms, from which existing species and varieties have

descended. Such a suggestion in the present writer's opinion is in one

sense a truism and in another an absurdity. That the potentiality of all

the characters of all the forms that have existed, pterodactyls,

dinosaurs, butterflies, birds, etc. etc., including the characters of all

the varieties of the human race and of human individuals, must have been

present in the primordial ancestral protoplasm, is a truism, for if the

possibility of such evolution did not exist, evolution would not have

taken place. But that every distinct hereditary character of man was

actually present as a Mendelian factor in the ancestral _Amoeba_, and that

man is merely a group of the whole complex of characters allowed to

produce real effects by the removal of a host of inhibiting factors, is

incredible. The truth is that biological processes are not within our

powers of conception as those of physics and chemistry are, and Bateson's

hypothesis is nothing but the old theory of preformation in ontogeny. Just

as the old embryologists conceived the adult individual to be contained

with all its organs to the most minute details within the protoplasm of

the fertilised ovum or one of the gametes, so the modern Mendelian,

because he is unable to conceive or to obtain the evidence of the gradual

development of a hereditary factor, conceives all the hereditary factors

of the whole animal kingdom packed in infinite complexity within the

protoplasm of the primordial living cells. That man is complex and

_Amoeba_ simple is merely a delusion; the truth according to Mendelism is

that man is merely a fragment of the complexity of the original _Amoeba_.

Mendelism studies especially the heredity of characters, and only

incidentally deals with recorded instances of the appearance of new forms,

such as the origin of a salmon-coloured variety of _Primula_ from a

crimson variety. The occurrence of new characters, or mutations as they

are called, has been specially studied by other investigators, and I

propose briefly to consider the two most important examples of such

research, namely, that by Professor T. H. Morgan, which deals with the

American fruit-fly _Drosophla_, and the other which concerns the mutations

of the genus of plants OEnothera, exemplified by our well-known Evening

Primrose.

Professor T. H. Morgan informs us [Footnote: _A Critique of the Theory of

Evolution_ (Oxford Univ. Press, 1916), p. 60] that within five or six

years in laboratory cultures of the fruit-fly, _Drosophila ampelophila_,

arose over a hundred and twenty-five new types whose origin was completely

known. The first of these which he mentions is that of eye colour,

differing in the two sexes, in the female dark eosin, in the male

yellowish eosin. Another mutation was a change of the third segment of the

thorax into a segment similar to the second. Normally the third segment

bears minute appendages which are the vestiges of the second pair of

wings; in the mutant the wings of the third segment are true wings though

imperfectly developed. A factor has also occurred which causes duplication

of the legs. Another mutation is loss of the eyes, but in different

individuals pieces of the eye may be present, and the variation is so wide

that it ranges from eyes which until carefully examined appear normal, to

the total absence of eyes. Wingless flies also arose by a single mutation.

These were found on mating with normal specimens to be all recessive

characters, thus agreeing with Bateson's views. The next one described is

dominant. A single male appeared with a narrow vertical red bar instead of

the broad red normal eye. When this male was bred with normal females all

the eyes of the offspring were narrower than the normal eye, though not so

narrow as in the abnormal male parent. It may be pointed out that this is

scarcely a sufficient proof of dominance. If the mutation were due to the

loss of one factor affecting the eye, the heterozygote carrying the normal

factor from the mother only might very well develop a somewhat imperfect

eye.

Morgan arranges the numerous mutations observed in _Drosophila_ in four

groups, corresponding in his opinion to the four pairs of chromosomes

occurring in the cells of the insect. After the meiotic or reduction

divisions each gamete of course contains in its nucleus four single

chromosomes. One of the four pairs consists of the sex-chromosomes. All

the factors of one group are contained in one chromosome, and it is found

in experiments that the members of each group tend to be inherited

together--that is to say, if two or more enter a cross together, in other

words, if a specimen possessing two or more mutations is crossed with

another in which they are absent, they tend to segregate as though they

were a single factor. This fact agrees with the hypothesis that the

factors in such a case are contained in a single chromosome which

segregates from the fellow of its pair in the reduction divisions.

Exceptions may occur, however, and these are explained by what is called

'crossing over.' When one chromosome of a pair, instead of being parallel

to the other in the gametocyte, crosses it at a point of contact, then

when the chromosomes separate, part of one chromosome remains connected

with the part of the other on the same side and the two parts separate as

a new chromosome, so that two factors originally in the same chromosome

may thus come to lie in different chromosomes. In consequence of this, two

or more factors which are usually 'coupled' or inherited together may come

to appear in different individuals.

Morgan emphasises the statement that a factor does not affect only one

particular organ or part of the body. It may have a chief effect in one

kind of organ, _e.g._ the wings or eyes, but usually affects several parts

of the body. Thus the factor that causes rudimentary wings also produces

sterility in females, general loss of vigour, and short hind legs.

The facts to which I shall refer concerning _Oenothera_ are for the most

part quoted on the authority of Dr. Ruggles Gates, and taken from his book

_The Mutation Factor in Evolution_ (London, 1915). The occurrence of

mutations in _Oenothera_ was first noticed by De Vries, the Dutch

botanist, in the neighbourhood of Amsterdam in 1886. He found a large

number of specimens of _Oenothera Lamarckiana_ growing in an abandoned

potato-field at Hilversum, and these plants showed an unusual amount of

variation. He transplanted nine young plants to the Botanic Garden of

Amsterdam, and cultivated them and their descendants for seven generations

in one experiment. Similar experiments have been made by himself and

others. The large majority of the plants produced from the _Oe.

Lamarckiana_ by self-fertilisation were of the same form with the same

characters, but a certain percentage presented 'mutations'--that is,

characters different from the parent form, and in some cases identical

with those of plants occurring occasionally among those growing wild in

the field where the observations began. Nine of these mutants have been

recognised and defined, and distinguished by different names. The

characters are precisely described and in many cases figured by Gates in

the volume cited above. The first mutant to be recognised--in 1887--was

one called _lata._ It must be explained that the young plant of

_Oenothera_ has practically no stem, but a number of leaves radiating in

all directions from the growing point which is near the surface of the

soil. The plant is normally biennial, and in the first season the

internodes are not developed. This first stage is called the 'rosette.'

From the reduced stem are afterwards developed one or more long stems with

elongated internodes, bearing leaves and flowers. In the mutation _lata_

the rosette leaves are shorter and more crinkled than those of

_Lamarckiana,_ and the tips of the leaves are very broad and rounded.

The stems of the mature plant are short and usually more or less decumbent

with irregular branches. The flower-buds are peculiarly stout and

barrel-shaped, with a protrusion on one side. The seed-capsules are

short and thick, containing relatively few seeds, and the pollen is

wholly or almost wholly sterile.

It is to be noted here, a fact emphasised by DeVries in his earliest

publications on the subject, that in nearly all, if not all cases, a

mutation does not consist in a peculiarity of a single organ, but in an

alteration of the whole plant in every part. In this respect mutations as

observed in _Oenothera_ seem to be in striking contrast to the majority of

Mendelian characters. Mutation in fact seems to be a case of what the

earlier Darwinians called correlation, while Mendelian characters may

apparently be separated and rejoined in any combination. For example, in

breeds of fowls any colour or any type of plumage may be obtained with

single comb or with rose comb. In my own experiments on fowls the loose

kind of plumage first known in the Silky fowl, which is white, could be

combined with the coloured plumage of the type known as black-red. At the

same time it must be borne in mind that since the factor, whether a

portion of a chromosome or not, is transmitted in heredity as a part of a

single cell, the gamete, and since every cell of the developed individual

is derived by division from the single zygote cell formed by the union of

the two gametes, the factor or determinant must be contained in every cell

of the soma, except in cases where differential division, or what is

called somatic segregation, takes place. Thus the factor which causes the

comb to be a rose comb in a fowl must be present in the cells that produce

the plumage or the toes or any other part of the body. Morgan, as

mentioned above, finds in _Drosophila_ that factors do affect several

parts of the body. It is, however, curious to consider that the factor

which produces intense pigmentation of the skin and all the connective

tissue in the Silky fowl has no effect on the colour of the plumage in

that breed, which is a recessive white. The plumage is an epidermic

structure, and therefore distinct from the connective tissue, but it is

difficult to understand why a pigment factor though present in every cell

has no effect on epidermic cells.

The Mendelians, when the mutations of _Oenothera_ were first described,

endeavoured to show that they were merely examples of the segregation of

factors from a heterozygous combination. They suggested in fact that

_Oenothera Lamarckiana_ was the result of a cross, or repeated crosses,

between plants differing in many factors, that the numerous mutations were

similar to the variety of different types which are produced by breeding

together the grey mice arising from a cross between an albino and a

Japanese waltzing mouse in Darbishire's experiment. Since that time,

however, the natural distribution and the cultural history of _Oenothera_

has been very thoroughly worked out. _Oenothera Lamarckiana_ is the common

Evening Primrose of English gardens. The species of the sub-genus _Onagra_

to which _Lamarckiana_ belongs were originally confined to America

(Canada, United States, and Mexico), but _Lamarckiana_ itself has never

been found there in a wild state. Attempts, however, to produce it by

crossing of other forms have not succeeded, and a specimen has been

discovered at the Muséum d'Histoire Naturelle at Paris, collected by

Michaux in North America about 1796, which agrees exactly with the

_Oenothera Lamarckiana_ naturalised or cultivated in Europe. The plant was

first described by Lamarck from plants grown in the gardens of the Muséum

d'Histoire Naturelle, under the name _OE. grandiflora_, which had been

introduced by Solander from Alabama, but Seringe subsequently decided that

Lamarck's species was distinct from _grandiflora_, and named it

_Lamarckiana_. Gates states that Michaux was in the habit of collecting

seeds with his specimens, and that it is therefore highly probable that

Lamarck's specimens were grown directly from seeds collected in America by

Michaux. Gates considers that the suggestion of the hybrid origin of

_Lamarckiana_ in culture is thus finally disposed of. By the year 1805,

_Lamarckiana_ was apparently naturalised and flourishing on the coast of

Lancashire, and in 1860 it was brought into commerce, probably from these

Lancashire plants, by Messrs, Carter. The cultures of De Vries are

descended from these commercial seeds, but the Swedish race of

_Lamarckiana_, as well as those of English gardens, differ in several

features and must have come from another source or been modified by

crossing with _grandiflora_. This last remark is quoted from Gates, but it

seems improbable that the Dutch plants should be derived from those of

Lancashire, and those of English gardens from a different source. The fact

seems to be, according to other parts of Gates's volume, that there are

various races of _Lamarckiana_ in English gardens and in the Isle of

Wight, as well as in Sweden, etc., and that these races differ from one

another less than the mutants of De Vries and his followers.

An important point about these mutations is that their production is a

constant feature of _Lamarckiana_. Whenever large numbers of the seeds of

this plant are grown, a certain proportion of the plants developed present

these _same_ mutations; not always all of them--some may be absent in one

culture, present in another, but four of them are fairly common and of

constant occurrence. The total proportion of mutant plants compared with

the normal was 1.55 per cent. in one family, 5.8 per cent. in another. It

would appear therefore, supposing that mutations arose subsequently in the

same determinate way from previous mutations, that evolution, though in a

number of divergent directions from one ancestral form, would proceed

along definite lines, and that there would be nothing accidental about it.

We should thus arrive at a demonstration of what Eimer called

orthogenesis, or evolution in definite directions.

The mutation _lata_ cannot be said to breed true, as the pollen is almost

entirely sterile. It has therefore been propagated by crossing with

_Lamarckiana_ pollen, with the result that both forms are obtained

with _lata_ varying in proportion from 4 per cent. to 45 per cent.

_Rubrinervis_ is a mutation from _Lamarckiana_, chiefly distinguished by

red midribs in the leaves and red stripes on the sepals. When propagated

from self-fertilised seed it produced about 95 per cent. of offspring with

the same characters, and the remaining 5 per cent. mutants, one of which

was _laevifolia_ which had been found by De Vries among plants growing

wild at Hilversum. Gates obtained a single plant among offspring of

_rubrinervis_ in which the sepals were red throughout, and to this he gave

the name _rubricalyx_. When selfed this plant gave rise to both

_rubricalyx_ and _rubrinervis_, and in the second generation when the

_rubricalyx_ was selfed again the numbers of the two were approximately 3

to 1. _Rubricalyx_ is therefore a dominant heterozygote, and this fact was

further confirmed in the third generation when a selfed plant gave 200

offspring all _rubricalyx_, the mother plant having evidently been

homozygous for the red character. In this case, therefore, we have what

Bateson was seeking, the origin of a new dominant character under

observation, the original mutation having arisen in a single gamete of the

zygote which gave rise to the plant. It is claimed by mutationists that

mutations are not new combinations or separations of Mendelian unit

characters already present, but are themselves new characters, though not

always necessarily, as in the case of _rubricalyx_, new unit characters in

the Mendelian sense.

Perhaps the most interesting of the researches on the phenomena of

mutation are those concerning the relation of the characters to the

chromosomes of the cell, in which Gates has been a pioneer and one of

the most industrious and successful investigators. The behaviour of

the chromosomes in meiosis or reduction division both in the pollen

mother-cells and in the megaspore mother-cells which give rise to the

so-called embryo-sac are fully described by Gates. Here it is only

necessary to refer to the abnormalities in the reduction division which

are related to mutation, and the results of these abnormalities in the

number of chromosomes. The original number of chromosomes in _OEnothera_

is 14. In the mutation _lata_ this has become 15, and also in another

mutation called _semilata_. The chromosomes before the reduction division

are arranged in pairs, each pair consisting, it is believed, of one

paternal and one maternal chromosome. One of each pair goes into one

daughter-cell and the other into the other, but not all maternal into one

and all paternal into the other. Thus each daughter-cell after the first

or heterotypic division in normal cases contains 7 chromosomes. A second

homotypic division takes place in which each chromosome splits into two as

in somatic divisions, and thus we have 4 gametes with 7 chromosomes each.

Now when _lata_ is produced it is believed that in the heterotypic

division one pair passes into one daughter-cell instead of one chromosome

of the pair into each daughter-cell, the other pairs segregating in the

usual way. We thus have one daughter-cell with 8 chromosomes and the other

with 6. This 6+8 distribution has actually been observed in the pollen

mother-cell in _rubrinervis_. When a gamete with 8 chromosomes unites in

fertilisation with a normal gamete with 7 the zygote has 15. The _lata_

mutants having an odd chromosome are almost completely male-sterile, and

their seed production is also much reduced: but this partial sterility

cannot be attributed entirely to the odd chromosome because _semilata_,

which has also 15 chromosomes, does not show the same degree of sterility.

Other cases occur in which the number of chromosomes in the somatic cells

is double the ordinary number--namely, 28--and others in which the number

is 21. The normal number in the gamete, 7, is considered the simple

or haploid number, and therefore the number 28 is called tetraploid.

This doubling of the somatic number of chromosomes is now known in a

number of plants and animals. It occurs in the _OEnothera_ mutant _gigas_.

The origin of it has not been clearly made out, but it must result either

from the splitting of each chromosome or from the omission of the

chromosome reduction. In many cases the more numerous chromosomes are

individually as large as those in normal plants, and consequently the

nucleus is larger, the cell is larger, and the whole plant is larger in

every part. But giantism may occur without tetraploidy, and vice versa. In

the _OEnothera gigas_ the rosette leaves are broadly lanceolate with

obtuse or rounded tips, more crinkled than in _Lamarckiana_, petioles

shorter. The stem-leaves are also larger, broader, thicker, more obtuse,

and more crinkled than in _Lamarckiana_. The stem is much stouter, almost

double as thick, but not taller because the upper internodes are shorter

and less numerous. It is difficult to avoid the conclusion that the

stouter character of the organs in this plant is causally connected with

the increased number of chromosomes. Where the number of cells formed is

approximately similar, as in two allied forms of plant in this case, the

greater size of the cells would naturally give a stouter habit, but it is

clear that large cells do not necessarily mean greater size. The cells of

_Salamander_ and _Proteus_ are the largest found among Vertebrates, but

those Amphibia are not the largest Vertebrates. It is curious to note how

different are these discoveries concerning differences in the _number_ of

chromosomes from the conception of Morgan that a mutation depends on a

factor situated in a part of one chromosome.

More copious details concerning mutations will be found in the

publications cited. The question to be considered here is how far the

claim is justified that the facts of this kind hitherto discovered afford

an explanation of the process of evolution. It seems probable that

mutations are of different kinds, as exemplified in _Oenothera_ by _gigas_

and _rubricalyx_ respectively, the former producing only sterile hybrids,

the latter behaving exactly like a Mendelian unit. There can be little

doubt that, as Bateson states, numerous forms recognised as species or

varieties in nature differ in the same way as the races or breeds of

cultivated organisms which differ by factors independently inherited.

There are facts, however, which prove that all species are not sterile

_inter se_, and that their characters when they are hybridised do not

always segregate in Mendelian fashion. John C. Phillips, [Footnote:

_Journ. Exper. Zool._, vol. xviii., 1915.] for example, crossed three wild

species of duck, _Anas boscas_ (the Mallard) with _Dafila acuta_ (the

Pintail) and with _Anas tristis_. In the former cross he states that

except for one or two characters there seemed to be no more tendency to

variation in the _F2_ generation than in the _F1_. An _F1_ Pintail-Mallard

[female] was mated with a wild Pintail [male]. According to Mendelian

expectation the offspring of this mating should have been half Pintail and

half Pintail-Mallard hybrids, but Phillips states that on casual

inspection the plumage of all the males appeared pure Pintail although the

shape was distinctly Mallard-like. The statement is, however, open to

criticism. The question is, what were the unit characters in the parent

species? If the unit characters were very small and numerous, an

individual in which all the characters of the Pintail existed together

among the offspring of the hybrid mated with pure Pintail would be rare in

proportion to the individuals presenting other combinations. Of the _F2_'s

obtained from crossing _Anas tristis_ [male] with _Anas boscas_ [female]

Phillips obtained 23 females and 16 males. The females were all alike and

similar to _F1_ females. Of the males one was a variate specially marked,

about half-way between the _F1_ type and the Mallard parent. This,

according to Phillips, was a segregate. The rest showed a range of

variation but no distinct segregation.

It is somewhat surprising that Mendelian experts, who seem to believe that

species are distinguished by Mendelian characters, have not made

systematic experiments on the crossing of species in order to prove or

disprove their belief.

For my own part I cannot help thinking that the origin of varieties in

species in a domesticated or cultivated state is in a sense pathological.

Such variation doubtless occurs in nature, but not with such luxuriance.

The breeds of domestic fowls differ so greatly that Bateson and others

refuse to believe that they have all arisen from the single species

_Gallus bankiva_. It seems to me from the evidence that there cannot be

any doubt that they have so arisen. One fact that impresses my mind is

that if we consider colour variations in domesticated animals, we find

that a similar set of colours has arisen in the most diverse kinds of

animals with sometimes certain markings or colours peculiar to one group,

_e.g._ dappling in horses, wing bars in pigeons. Thus in various kinds of

Mammals and Birds we have white and black, red or yellow, chocolate with

various degrees of dilution, and piebald combinations. Why should forms

originally so different, as the cat with its striped markings and the

rabbit with no markings at all, give rise to the same colour varieties? It

seems probable that the reason is that the original form had the small

number of pigments which occur mixed together in very small particles, and

that in the descendants the single pigments have separated out, with

increase or decrease in different cases. It is true that historical

evidence tends to show that the greatest variations, such as albinism in

one direction or excess of pigment in the other in the Sweet Pea, were the

first to arise (see Bateson, Presidential Address to British Association,

Australia, 1914, Part I.), and the splitting appears often to be

intentionally produced by crossing these extreme variations with the

original form, but the possibility remains that the conditions of

domestication, abundant food, security and reduced activity, lead to

irregularity in the process of heredity. In any case the mere separation

among different individuals of factors originally inherited together in

one complex does not account for the origin of the complex or of the

factors. This is somewhat the same idea as that of Bateson when he states

that it is easy to understand the origin of a recessive character but

difficult to conceive the origin of a dominant.