Popular Science Monthly/Volume 62/January 1903/Variation in Man and Woman


ARE variations more common in males than in females? That is a question which has passed through various phases during the past century. John Hunter, who touched on the matter from a biological standpoint, vaguely indicated that males are more variable than females. Meckel, on the contrary, came to the conclusion, on pathological grounds, that in the human species females show a greater degree of variability, and he thought that since man is the superior animal and variation a sign of inferiority the conclusion was justified. "We may state as a principle," Meckel wrote ninety years ago at the outset of his manual of descriptive and pathological anatomy, "that anomalies are more common in the female. This phenomenon seems to depend on the eighth law [Meckel's 'law of development,' according to which woman is more primitive than man] since the organization of the female results from development being arrested at an inferior degree." But while he regards deviations as on the whole more common in woman he admits certain exceptions, and more especially instances the heart and the bladder as more variable in man. Meckel was a profound student of anatomy, but not a very luminous thinker. Some years later Burdach took up the question in his 'Physiologie.' That great biologist at once raised the problem to a higher level, realized its wider bearings and cleared away the prejudices which had surrounded it. He recognized that in some respects women are more variable than men, but pointed out that, contrary to Meckel's opinion, this was no indication of woman's organic inferiority. He showed from the statistics of the Anatomical Institute of Konigsberg that we must distinguish between different kinds of abnormality. Further he referred to the facts that indicate that woman is more childlike than man, but, he added, "it is a very common but a very gross error to consider age as a scale of perfection and to regard the child as absolutely imperfect as compared to the adult. It is not imperfection but simply certain childlike characteristics which women preserve"; and, he points out, it is in decrepitude that women take on the characteristics of the so-called superior sex. His general conclusion was that the nature of man and the nature of woman are both excellent, but there are wider variations in men, more genius and more idiocy, more virtue and more vice.

Darwin turned his attention towards this point and accumulated data. In the 'Descent of Man' he brought together many of the chief facts then known concerning variation in man and woman. All the evidence that he could find pointed in the same direction, and he concluded (Part II., Ch. 8) that there is a 'greater general variability in the male sex.'

Some twenty years later in a summary study of human secondary sexual characters entitled 'Man and Woman,' written as a brief introduction to a more elaborate study of the sexual instinct in man, I devoted a chapter to this question, dealing with it more comprehensively than had previously been done and drawing data from a much wider field, but finding no reason to differ fundamentally from the conclusions of Hunter, Burdach and Darwin. I could not indeed assert that as regards man the greater variability of the male is 'general,' but all the facts available since Darwin's day indicated that a greater variability of the male occurred in the majority of the groups of data investigated. And when I considered that this greater organic variational tendency of men is apparently true of psychic variations also—of genius, of idiocy and other mental anomalies having an organic basis—it seemed to me that in the greater variational tendency of man we are in the presence of a fact that has social and practical consequences of the widest significance, a fact which has affected the whole of our human civilization. Although the greater variational tendency of men is balanced by the more equable level of women, we have to recognize that the existence of the exceptional men who have largely created the lines of our progress is based on natural law. It is a conclusion which does not yet appear to me to be fundamentally affected.

There was, however, one important omission in my statement of this question, and I wish to emphasize the importance of the omission because its significance will subsequently become apparent to the reader. I said little or nothing as to the variability of men and women in size, either as regards total stature and weight, or the dimensions of parts of the body.[1] The reason for that omission is clearly indicated in various parts of the volume and we shall encounter it in due course.

Three years later, in a volume of miscellaneous essays entitled 'The Chances of Death,' Professor Karl Pearson published a lengthy paper entitled 'Variation in Man and Woman.' This writer started with the assertion that in 'Man and Woman' I had 'done much to perpetuate some of the worst of the pseudo-scientific superstitions, notably that of the greater variability of the male human being' and that it was the object of his essay 'to lay the axe to the root of this pseudo-scientific superstition.' In fact, as he is careful to tell us at frequent intervals, before he himself entered the field (a field, be it remembered, occupied by some of the world's greatest biologists) all was 'dogma,' 'superstition,' 'nearly all partisan,' at the best 'quite unproven,' I am inclined to think that these terms, which spring so easily to Mr. Pearson's pen, are automatic reminiscences of the ancient controversies he has waged with theologians and metaphysicians. They are certainly a little out of place on the present occasion.

In selecting the material for his demonstration, Professor Pearson tells us he sought to eliminate all those 'organs or characteristics which are themselves characteristic of sex,' such being, in his opinion, gout and color-blindness; he also threw aside all variations which can be regarded as 'pathological,' on the hypothetical ground that such 'pathological' variations may have a totally different sexual distribution from 'normal' variations. He decided that size is the best criterion of variability. As to how a 'variation' may be defined Professor Pearson makes no critical inquiry, though such inquiry would very seriously have modified his final conclusions.[2]

"What we have to do," he states, "is to take healthy normal populations of men and women, and in these populations measure the size of organs which do not appear to be secondary sexual characters, or from which the sexual character can be eliminated by dealing solely with ratios." Various kinds of size are therefore selected for treatment, such as that of the skull, chiefly as regards its capacity and length-breadth index, stature, span, chest-girth, weight of body and of various internal organs, etc., all these, it is observed, being various aspects of the one factor of size. It is shown by careful treatment of the available data the so-called coefficient of variation being accepted as a possible or indeed probable measure of significant variation—that, as far as there is any difference at all, women are, on the whole, slightly more variable than men. Having reached this result the author leaps bravely to the conclusion, that 'accordingly, the principle that man is more variable than woman must be put on one side as a pseudo-scientific superstition.'

If a reply has so far not been forthcoming from the writer against whom this elaborate paper was chiefly directed, this has not been either because I admitted the justice of its conclusions, or complacently accepted a damnation to which I had been consigned in very excellent company. The subject lay only on the outskirts of my own field; I could claim no originality in it; all that I had done was to sift and bring to a focus data which had hitherto been scattered, and to show their significance. At the same time I again placed the subject on my agenda paper for reconsideration. In the meanwhile, it scarcely appeared that Mr. Pearson's arguments met with much acceptance, even from those whom they most concerned.[3]

Almost the only attempt to consider them, indeed, which I have met, is in a review of 'The Chances of Death' by Professor W. F. II. Weldon, in Natural Science. This sympathetic critic, with a biologist's instincts, clearly felt that there was something wrong with Mr. Pearson's triumphant demonstration, although as the subject lay outside his own department he was not able to indicate the chief flaws.

There is indeed one initial flaw in Professor Pearson's argument, to which Professor Weldon called attention; it could scarcely fail to attract the notice of a biologist. We are told that we must put aside 'characteristics which are themselves characteristics of sex,' like gout and color-blindness, which 'without being confined to one sex' are yet peculiarly frequent in one sex. Thus, we see, characteristics not confined to one sex may yet be characteristic of one sex, and when we seek to find what characters vary more in one sex than in the other, we must carefully leave out of account all these characters which are most clearly more prevalent in one sex. Professor Pearson thus sets out with an initial confusion which is never cleared up. His object, he tells us, is to seek such degrees of variability as are 'secondary sexual characters of human beings,' and we infer from the course of his argument that the desired characters while not confined to one sex must yet be peculiarly frequent in one sex. Yet these are precisely the group of characters ruled inadmissible at the outset! No definition of secondary sexual characters is anywhere given, or on such premises could be given.

Professor Pearson seems to assume that the conception of a secondary sexual character is too obvious to need definition. As a matter of fact there is considerable difference of opinion. Since Hunter first spoke of the 'secondary properties' of sex, which he regarded as dependent on the primary, only developing at puberty, and principally, though not entirely, confined to the male, the conception has very much changed; there has been a tendency to throw all sorts of miscellaneous sexual differences into the category. I have suggested that it would be convenient to introduce a group of 'tertiary' sexual characters, keeping the term 'secondary' to its original sense and reserving as 'tertiary' all those minor differences which are not obvious, which can have no direct influence on mating, and only exist as averages; such are the composition of the blood and the shape of the bones.[4]

It is difficult to correct all the errors and confusions which Professor Pearson falls into at this point. He remarks that we must not regard the greater prevalence of idiocy among men as evidence of greater male variability, unless we count on the other side the greater prevalence of insanity among women. The error here is double. As a matter of fact, although in England and Wales during recent years the incidence of insanity has been as great on women as on men,[5] nearly everywhere else it is markedly greater in the case of men. Indeed even in England and Wales, at the present time, if we may trust the Commissioners in Lunacy in their latest annual report (1902), the incidence of insanity as indicated by the admissions to asylums, is, in ratio to the male population, still slightly greater in the case of men. Professor Pearson has been misled by the greater accumulation of females in asylums, failing to take into consideration the greater longevity of women, which among the insane is specially marked.[6] But even if the facts had been as stated by Mr. Pearson, his inference would still have been wrong; idiocy is mainly a congenital condition and therefore a fairly good test of organic variational tendency; insanity, though usually on a hereditary basis, is invariably an acquired condition, dependent on all sorts of environmental influences, so that it can not possibly furnish an equally fundamental test. Color-blindness, Mr. Pearson also tells us, is a peculiarly male 'disease,' and must not be used as an argument for greater variability in men unless we use the prevalence of cancer of the breast in women on the other side. Again there is a double error; not only is a congenital anomaly improperly compared to an acquired disease, but a gland like the breast which is only functional in one sex is paired off with an organ like the eye which is equally functional in both sexes. The prevalence of gout among men is, again, paired off against the prevalence of hysteria among women. Here the error is still more complex. Not only is gout not a truly congenital condition, though, like insanity, it frequently has a hereditary basis, but if we take into account conditions of 'suppressed' gout it is by no means more prevalent in men than in women, and even if we do not take such conditions into account, it is still not possible to pair off gout against hysteria, since, although in some countries hysteria is more prevalent in women, in others (as, according to some of the best authorities, in France) it is found more prevalent in men. But it would be tedious to explore further this confused jungle of misstatements.

From the point of view of sexual differences in variational tendency it is not necessary to exclude rigidly either 'tertiary,' 'secondary,' or even 'primary' sexual characters, provided we are careful to avoid fallacies which are fairly obvious, and do not compare organs and characters which are not truly comparable. Even those secondary sexual characters which are almost or entirely confined to one sex may properly be allowed a certain amount of weight as evidence, especially if we grant that such characters are merely the perpetuation of congenital variations. If, therefore, as is generally agreed, such characters more often occur in males, that fact is a presumption on the side of a greater male variational tendency which there is no reason entirely to ignore. It is not conclusive, but it must receive its due weight. To assume, with Professor Pearson, that a variation has no variational significance because it occurs often in one sex and seldom in the other[7] seems altogether unwarrantable.

If, however, Professor Pearson's attempt to discriminate between different kinds of sexual characters from the point of view of sexual variability fails to work out, and is in any case unnecessary, at another point he falls into the opposite mistake of making no attempt to discriminate when discrimination is of the first importance. As we have already incidentally seen, it seems to him to be of no importance whether the variational tendency is tested by variations having an organic congenital base, or by variations which may be merely due to environmental influences during life.[8] To him they are all alike i variations, 'and the most important are those that can most conveniently be caught in the mathematical net. Indeed he goes further than this. He actually discriminates against the more organic and fundamental kinds of variation. It seems to him 'erroneous' to take into account congenital abnormalities of any kind when we wish to test the relative variability of the sexes. In determining the variational tendencies of the sexes we must leave out of account the majority of variations!

The ground on which Professor Pearson rejects abnormalities is that they are 'pathological,' and that it is conceivable that pathological variation might be greater and normal variation less in the same sex.[9] He believes that in regarding the 'normal' and the 'abnormal' as two altogether different and possibly opposed groups of phenomena he is warranted by 'current medical science.'

This is very far indeed from being the case. It is quite true that in ordinary clinical work the physician does make such a distinction; it is practically convenient. But it is not science, and if the physician is a genuine pathologist he admits that it is not. This is so well recognized that I had thought it sufficient to quote the remark of the greatest of pathologists, Virchow, to the effect that every deviation from the parental type has its foundation in a pathological accident—a statement which Professor Pearson, on the strength of what is really a verbal quibble, contemptuously puts aside as 'meaningless.' We ought not to say the 'parental type,' he tells us, we ought to say 'a type lying between the parental type and the race type'; let us say it—and the statement remains substantially the same, so far as the question before us is concerned.[10]

Virchow is by no means the only pathologist of high authority who has distinctly laid down this principle. Thus, as Ballantyne points out—when remarking that the ancient belief, held even by Simpson, that anomalies and malformations are due to disease has been supplemented by modern research—Mathias Duval has emphatically declared that it is not to be thought that the malformation of any part is a result of disease of that part.[11]

Even, however, if we go back to the time of Simpson, and earlier, we find that Meckel—who is sometimes regarded as one of the founders of the study of variations—clearly recognized that the simplest anomalies and varieties pass gradually into monstrosities, and that the same laws apply to both.[12]

Indeed so did Hunter in the previous century. 'Every deviation' he wrote at the outset of his almost epoch-marking 'Account of an Extraordinary Pheasant,' 'may not improperly be called monstrous,' so that 'the variety of monsters will be almost infinite.'

The tendency of scientific pathology is at once to push the frontiers of the normal into regions popularly regarded as belonging to disease, and at the same time, when actual disease comes into question, to refuse to admit that any new laws are brought into operation. "Between any form of disease and health," one of the founders of modern pathology declared a quarter of a century ago, "there are only differences of degree. No disease is anything more than an exaggeration or disproportion or disharmony of normal phenomena."[13] The notion that disease and health are distinct principles or entities Bernard regarded as a sort of idea belonging to the medical lumber

room. These conceptions have been brilliantly developed in the work of recent pathologists.[14]

And if it is argued that a mathematician cannot be supposed familiar with the principles of pathology, it must be replied that Mr. Pearson has here ventured along a path which leads immediately up to these principles, and, further, that the principle in question is so simple and elementary that it may already be said to have entered general culture. I take up the latest volume of Nietzsche's works ('Der Wille zur Macht')—written more than ten years ago, though only now published—and read: "The value of all morbid conditions is that they show us in magnified form certain conditions that are normal, but in the normal condition not easily visible. . . . Health and disease are not essentially different, as the old physicians and some modern practitioners have believed. To regard them as distinct principles struggling for the living organism is foolish nonsense and chatter."

On the whole, then, there is no reason for rejecting abnormalities when we are considering the relative variational tendencies of men and women. To the mathematical mind—Professor Pearson forces us to admit—it is possible to conceive that the laws of pathology may reverse the laws of physiology, but such a conception the biologist regards as absurd.

More than this must, however, be said. Not only can we not leave anomalies out of account in dealing with this question, but it is precisely the anomalies which furnish us with the most reliable evidence. The word 'abnormality' is apt to mislead, and Professor Pearson somewhat prejudices the matter in unscientific ears by insisting on its use. It is not a scientific term; the so-called anomaly is not abnormal in the sense that it is morbid; it is only exceptional. It merely indicates the extreme swings of a pendulum whose more frequent oscillations are popularly regarded as 'normal.' What is commonly termed an 'anomaly' might really be regarded as the 'variation' par excellence.

Such an assertion would be by no means arbitrary. It does in fact correspond with the usage of most of the writers who have investigated this matter until the present day, and it is possible to justify such usage. If—to return to the image of the pendulum—we wish to find out whether the male or female pendulums swing farthest, we must so far as possible let them swing freely; the more they are restrained by external forces the less the significance of the results we reach. Now the congenital 'anomalies' are precisely the kind of variation that most nearly corresponds to the free swing of the pendulum. It is true that there is no absolute distinction between the initial energy and the subsequent modifying influences, but it is equally true that if we wish to measure and compare the aboriginal energies of the male and female organisms, we must so far as possible disregard those characters which are very considerably influenced by late modifying forces.

Professor Pearson has, however, chosen, as a final and crucial test of the variational tendency in men and women, the single point of difference in size, chiefly in adults. That is to say, he has selected as a final and unimpeachable test one of the most fragile of distinctions, a distinction that has been exposed to a lifetime of modifying influences that are incalculable.

Even if we admit that size at birth constitutes a sound test and this can not be admitted without qualification, as we shall soon see—it is evident that the comparative variation of the sexes in this respect is liable to be affected by environmental circumstances as age increases. The influences of life differently affecting and exercising the two sexes, the influence of death probably exerting an unequal selective influence—both alike must be allowed for if this kind of evidence is to be regarded as a test of the first rank of importance. Otherwise we are not dealing with the incidence of variations at all, but with the elimination of variations—an altogether different matter. Professor Pearson himself gradually awakes to a realization of this fact as he proceeds with his task, and remarks at last that he strongly suspects that the slightly greater variability of woman which his results show is mainly due to a relatively less severe struggle for existence! Probably he is right, but if so his whole argument falls to the ground. The question of the organic variational tendencies of men and women remains untouched; we have been introduced instead to a problem in selection. So true is it that, as Bacon said, the half of knowledge lies in asking the right question.

We are bound to suppose that when Professor Pearson set out lie intended to use the term 'variation' in the same sense as his predecessors had used it—for otherwise his results could not validly be opposed to theirs—but it would appear that as he went on, by an unconscious process of auto-suggestion, he insensibly glided into a familiar field.[15]

It may seem unnecessary to pursue Professor Pearson any further. It is sufficiently clear that the inquiry he has carried out, however valuable it may be in other respects, has no decisive bearing on the question he undertook to answer, and can have no very damaging effect on the writers he attacks. But there is considerable interest in driving the point of the discussion still further home.

It may be agreed that since differences in size are probably affected by the influences of life and death to a considerable extent, and perhaps unequally eliminated in the two sexes, they do not form a reliable guide to the sexual incidence of variations. But, it may be argued, this cannot affect measurements made at birth, and we must therefore accept the validity of those of Professor Pearson's measurements which concern the infant at birth. Here, however, we encounter a fact which is of the first importance in its bearing on our subject: the elimination of variations in size has already begun at birth, and there is reason to suppose that that elimination unequally affects males and females. This was duly allowed for in 'Man and Woman,' but there is no hint of it throughout Professor Pearson's long paper. He does not dispute this influence, nor does he realize that until he has disputed it his conclusions can not be brought to bear against mine. Professor Pearson's earlier statistical excursions into the biological field were chiefly concerned with crabs; in passing from crabs to human beings he failed to allow for the fact that human beings do not come into the world under the same conditions. I make no large claim for superior insight in this matter; it was probably a question of training; I was practically familiar with the phenomena of childbirth; he was not. But his ignorance has profoundly affected the validity of his cherished criterion of sex variability, in so far as it is used against his predecessors in this field.

Every child who is born into the world undergoes a severe ordeal, due largely to the limited elasticity of the bony pelvic ring through which it has to pass. Probably as a result of this, a certain proportion perish as they enter the world or very shortly after. Among the number thus eliminated there appears to be a very considerable proportion of the largest infants. Doubtless because male infants tend to be larger than female infants, males suffer most at and shortly after birth. This appears to be the rule everywhere.[16]

So far as I am aware, the first attempt to explain this matter scientifically was made in 1786 by an English doctor named Clarke, physician to the Lying-in Hospital at Dublin.[17] By weighing and measuring 120 newborn infants of both sexes he found that there was a marked tendency for the males to be larger than the females. 'Hence appears' as he is pleased to put it, 'the merciful dispensations of Providence towards the female sex, for when deviations from the medium standard occur it is remarkable that they are much more frequently below than above this standard.' He considered that the greater mortality of males at and shortly after birth is largely due to the injuries to the head occurring at birth, but also that, since the males are larger and therefore make from the first a larger demand on the nutritive capacity of the mother, they are more likely to suffer from any defect of the mother in this respect. The problem and its possible and probable explanations were thus clearly stated more than a century ago.

As often happens with pioneers, Clarke's little paper was forgotten, and for more than half a century, although a number of workers brought extensive contributions of new data, their attitude was frequently illogical or one-sided, and the progress of scientific knowledge was not great. In 1844 Simpson published a well-known study which brought together a mass of evidence bearing more or less on the question before us. He showed that in male births the mothers suffered excessively as well as the infants; he refused to admit that the greater mortality of males at and shortly after birth could be due to any other cause than the generally recognized larger size of the male head (mainly on the ground that fœtal deaths up to birth are fairly apportioned to the two sexes) and concluded that the greater size of the male head is the cause of a vast annual mortality. A number of later obstetrical inquirers furnished additional contributions to the matter, at one point or another, though not always agreeing that so great a mortality could be due to a difference of size which seemed so small. One authority, indeed, roundly declared that the belief in the larger size of the male head was merely 'a popular prejudice'; this led to fresh measurements, and in this field Stadtfeldt of Copenhagen received credit which really belonged to Clarke of Dublin. Veit showed that even at equal weights more boys than girls die at birth, but, on the other hand, according to Pfannkuch's results, even at equal weights boys' heads are larger than girls'. In any case it certainly seemed probable that the larger size of the male child's head was an important factor in this mortality, and when at length the question began to attract the attention of statistical anthropologists this conclusion was confirmed. The Anthropometric Committee of the British Association, presided over by Mr. Francis Galton, in its final report in 1883 stated its belief that "it would appear that the physical (and most probably the mental) proportions of a race, and their uniformity within certain limits, are largely dependent on the size of the female pelvis, which acts as a gauge, as it were, of the race, and eliminates the largest infants, especially those with large heads (and presumably more brains), by preventing their survival at birth."

It must be added, however, that no direct and final demonstration has been brought forward of the tendency to the elimination of the males (or even infants independently of sex) of the greatest weight or those having the largest heads. For this we require to compare male and female stillborn infants at full term with those who are born living and which subsequently survived for at least a week (a longer period would be more desirable but difficult to secure). Such measurements are not to be found in medical literature, so far as I can discover; at the most we find averages, which are meaningless from the present point of view. I applied to obstetrical and anatomical authorities in various countries and received a number of interesting letters and data, including series of entries from the registers of maternity hospitals. But none of the series so far received contains a sufficient number of stillborn children. So far as they go, they are confirmatory of the belief that it is more especially the large children that are eliminated by the selection of birth. The largest series (with 60 stillborn male babies and 50 stillborn females), for which I am indebted to Dr. C. M. Green, of the Boston Lying-in Hospital, shows that among the stillborn of either sex the range of variation is greater than among the living of the same sex, the absolute range of variation being not only greater as compared with the living babies of the same sex, but there being a greater piling up at each end in the case of the stillborn. The data do not suffice to indicate that there is a greater mortality of the largest sized males than of the largest sized females, when we compare the stillborn with the living of the same sex and weight. Another series, more elaborate in its details but still smaller in number as regards the stillborn—for which I am indebted to Professor Whitridge Williams, of Johns Hopkins Hospital—leads to a similarly incomplete conclusion. I still await more extensive data which have been promised me from a British source.[18]

There is, however, another test which, while it can by no means be put forward as having any statistical validity, yet furnishes a highly significant indication in this matter. Just as on the psychic side certain very rare individuals appear in the world whose intellectual capacity enormously excels that of their fellows, so, corresponding to 'genius,' we have on the physical side certain equally rare individuals who at birth enormously excel their fellows in physical size, while yet remaining normal and well proportioned. Now, we may ask, do these individuals possessing congenital physical 'genius' resemble persons of psychic genius in being more often male than female? Ordinary statistics are not here available, for these cases are so rare that they very seldom fall into an ordinary series. Smellic found one child weighing over 13 pounds in 8,000 cases; in France, a child of 12 pounds was only found in 20,000 cases.[19] As even a child of 9 pounds is generally considered large, it is clear that when we get beyond 13 pounds we reach a point at which the average difference between males and females is trifling, so that there is almost as great a chance of females as of males reaching the extremely large weights. The only practicable way of obtaining information concerning these cases lay in collecting the scattered records. I have collected all that I can find in medical journals of standing, chiefly English, during the past half century, being aided by the references in Neale's 'Medical Digest.' I have only noted the cases that appear to have been healthy and well developed and weighed over 13 pounds at birth. One unexpected difficulty I encountered: in many cases, even when numerous measurements were given, no reference was made to sex. While such cases were necessarily rejected, I may say that I think it probable that most, and perhaps all, of these rejected cases were males; this was so in the only case in which, by writing to the medical reporter immediately on publication, I was able to repair the omission; the medical mind seems to share in some degree the instinctive conviction that the typical human being is a male, and that in the case of males it is unnecessary to make any reference to sex. My cases were thus reduced to 21. Of these there were only 3 females to 18 males. The females all died at birth, as well as about half the males. However rough this method of estimation may be, it is highly improbable that any more methodical inquiry on children of this size would entirely reverse so large a preponderance of males.

Such a result, it will be seen, can not be considered as absolutely conclusive proof that there exists a selection of birth which in its operation tends to the destruction of the larger male children either at the moment of birth or during the succeeding days and weeks, though it renders such selection probable. This element of doubt, however, by no means makes Professor Pearson's position any stronger. It is sufficient to show that for more than a century past evidence has accumulated which indicates that the group of data on which Professor Pearson solely and absolutely relies for the foundation of his argument is modified by an influence which renders it tainted for such a purpose. In view of this circumstance, and of the fact that I had rejected this group of evidence on these grounds, the onus probandi clearly rested with Professor Pearson. In other words, he had to show either that male children are not larger than female children at birth, or else that large children do not suffer more than smaller children in passing through the maternal pelvis. The fact that Professor Pearson gives no indication that he had realized the necessity of this preliminary step is sufficient proof that he was not adequately equipped for the task he has undertaken.

We now come to a point which is not the less interesting for being entirely hidden from Mr. Pearson. It has been seen that the selection exercised by the pelvis to the detriment of male children is not absolutely proved. But if for the moment we assume that it exists, what are the phenomena that we should expect to find, as regards size, among the survivors? Obviously, a more or less diminished sexual difference during life, with a maximum of sexual difference immediately after birth[20] Now this is exactly what Professor Pearson found! 'Summing up in general our conclusions for weight,' he states, 'it would appear that, except at birth, man is not more variable than woman.' The very great significance of this exception, as affecting any argument on these premises brought against the position maintained in 'Man and Woman,' he undoubtedly failed to see. Still the exception evidently puzzled him. He accumulated series of data on the subject, and indeed initiated an entirely new and very extensive investigation. But the conclusion remained on the whole unaffected. Thus we see that our author, in all innocence, supplies a valuable piece of proof in favor of that very position which he imagines that he is upsetting! If this is the way that the axe is to be laid to the root of 'pseudo-scientific superstitions' they will certainly continue to flourish exceedingly.

We have now reached the climax of Professor Pearson's argument. It is from this giddy height that Mr. Pearson surveys with contempt those foolish persons who still believe that the variational tendency is greater in men than in women, and nothing further remains to be said. If instead of hastening to execute a war-dance on what he vainly imagined was the body of a prostrate foe, Mr. Pearson had pointed out, as he would have been quite warranted in doing, that his conclusions, so far as they rested on a definite basis of fact, confirmed the thesis maintained by Darwin and more fully enforced in 'Man and Woman,' his position would have been unimpeachable. If, again, he had refrained altogether from attempting to interpret his own data—a task for which, it is obvious, he was singularly ill-prepared—and had put them forth simply as a study in natural selection—which is what they really are—his position would, again, have been altogether justifiable. But as the matter stands he has enmeshed himself in a tangle of misapprehensions, confusions and errors from which it must be very difficult to extricate him.

It may be well to summarize briefly the main points set forth in the foregoing pages.

1. In opposition to the doctrine of Darwin, more fully set forth in my 'Man and Woman,' that the variational tendency is, on the whole, more marked in men than in women, Professor Pearson resolved to show that this is one of 'the worst of the pseudo-scientific superstitions.'

2. Unfortunately, however, it never occurred to him to define what ho meant by 'variation,' nor to ascertain what the writers whom he was opposing meant by the term.[21] A very little consideration suffices to show that a typical variation, in what may fairly be called its classic sense, is a congenital organic character on which selection works, while, as understood by Professor Pearson, though without definite statement, a typical variation is a character—of almost any kind, occurring at any period of life—produced by selection. 'To the biometrician,' Professor Pearson has recently stated, 'variation is a quantity determined by the class or group without reference to its ancestry.' That is to say, it need not be organic or congenital, and it must usually be modified, and sometimes entirely produced, by its environment. This definition may be better than the more classical conception of a variation. But it is certainly very different. To suppose that conclusions reached concerning this kind of variation can be used to overthrow conclusions reached concerning the other kind is obviously unreasonable.

3. Having silently adopted this conception of a variation, Professor Pearson proceeds to inquire what 'different degrees of variability are secondary sexual characters' and not 'characteristics which are themselves characteristics of sex'; and is hereby led into various eccentricities of assertion which it is unnecessary to recapitulate. 'Secondary sexual characters' remain, in his hands, like variation, undefined.

4. All 'abnormalities' are added to the material rejected as unsuitable for investigation, on the ground that they are 'pathological.' It has been easy to show that this notion cannot be maintained, and that in his pious horror of 'pseudo-scientific superstitions' Professor Pearson here lays himself open to retort. Anomalies are not pathological, except in the sense of Virchow, who regarded pathology as simply the science of anomalies. Moreover, scientific pathologists do not admit that even diseases can be regarded as involving any new or different laws. Morbid as well as normal phenomena alike furnish proper material, if intelligently used, for the investigation of this question.

5. Professor Pearson decides that differences in size furnish the best measure of the variability of the sexes. In reaching this decision he makes no reference to the fact that the probabilities accumulated during a century tended to discredit this group of evidence for the purposes he had in view.

6. If, however, we put aside those probabilities which tend to render this evidence tainted, so far as the object of Professor Pearson's special argument is concerned, we still find that the results he reaches are precisely the results we should expect if the position he assails is sound. That is to say that at birth, before the results of the assumed selective action of the pelvis have yet been fully shown, there is greater variability of the males, while later, as a result of that selection, there is a tendency to equality in sexual variability.

7. The net outcome of Professor Pearson's paper is thus found to be a confirmation of that very doctrine of the greater variational tendency of the male which he set out to prove to be 'either a dogma or a superstition.'

It may be as well to state, finally, that nothing I have said can be construed as an attempt to disparage those 'biometrical' methods of advancing biology of which Professor Pearson is to-day the most brilliant and conspicuous champion. I am not competent to judge of the mathematical validity of such methods, but so far as I am able to follow them I gladly recognize that they constitute a very valuable instrument for biological progress. I say nothing against the instrument: I merely point out that, on this occasion, the results obtained by its application have been wrongly interpreted.

  1. I did not consider that such evidence must be absolutely rejected—I admitted it in one or two cases (printed in smaller type)—but simply that as it was liable to a discount of unknown extent it could not be placed in the first rank of evidence.
  2. It is true, indeed, that Mr. Pearson remarks that the question 'What are the most suitable organs or characteristics for measuring the relative variability of man and woman?' 'really involves a definition of variability.' But he adds that 'the definition given may be so vague as to beg off-hand the solution of the problem we propose to discuss.' That suspicion, as we shall see, is not altogether unjustified.
  3. Mr. Pearson has endeavored to find an opponent of the greater variational tendency of men in Tennyson, who wrote:

    For men at most differ as Heaven and earth,
    But women, worst and best, as Heaven and Hell.

    This argument, however—whatever it may be worth—had already been answered by anticipation in a chapter of 'Man and Woman' on the affectability of woman in which I pointed out that the 'Heaven and Hell' of woman are both aspects of her greater affectability; not only does one woman differ from another as 'Heaven and Hell' but the same woman may so vary at different times.

  4. Professor Waldeyer, who has done me the honor of critically examining some of the main points in my book (in the form of an address at one of the annual meetings of the German Anthropological Society) is inclined to doubt the value of this distinction since there is no clear line of demarcation between secondary and tertiary characters. That, however, I had myself pointed out, and the objection cannot logically be held by any one who accepts secondary sexual characters and recognizes that they merge into the primary. There are many natural groups which have nuclei but no definite boundary walls.
  5. It so happens that I was the first to call attention to the fact that towards the end of the last century the number of women admitted to asylums in England and Wales had for the first time begun to exceed the number of men. (Art: 'Influence of Sex on Insanity,' in Tuke's 'Dictionary of Psychological Medicine.')
  6. How serious this fallacy is may be indicated by an illustration that chances to come to hand almost as I write. I read in a South American medical journal that in the Asylum of Santiago in Chili on the 1st of January, 1901, there were 560 men and 655 women, but, during the year, 539 men were admitted and only 351 women.
  7. In the case of ordinary gout, which Professor Pearson regards as typical of the tests to be excluded, opinions differ considerably as to sexual liability; according to one leading authority it is 68 men to 12 women.
  8. It is scarcely necessary to remark that the two groups cannot be absolutely separated.
  9. This conception, Mr. Pearson remarks, seems never to have occurred to me. In that shape, happily, it has not. But in 'Man and Woman' and elsewhere I have repeatedly called attention to the fact that, as regards various psychic and nervous conditions, while gross variations are more frequent in men, minor variations are more common in women. This seems to cover whatever truth there may be in Mr. Pearson's supposition.
  10. Virchow repeatedly emphasized the statement in question and by no means always in the form that offends Professor Pearson. Thus he remarked in 1894, at the annual meeting of the German Anthropological Society, that whenever 'the physiological norm hitherto subsisting is changed' we are in the presence of an anomaly and that in this sense every departure from the norm is a pathological event, though it is not a disease and may not be harmful, may even be advantageous. In what was perhaps his last utterance on the subject (Zeitschrift für Ethnologie, 1901, p. 213) he repeated that pathology as well as physiology is an essential factor in the development of the human race. Pathologists will, I know, agree with me that a conviction of the essential unity of physiology and pathology lay at the foundation of the pathological revolution which Virchow effected.
  11. With this result Dr. Ballantyne—who may be said to be the chief British authority on pathology in its antenatal aspects—in the main concurs. He even goes so far as to assert ('Manual of Antenatal Pathology,' 1902, p. 35) that natural birth in its effects on the child may almost be regarded as a pathological process; 'it is very certain that the same amount of distortion of parts, occurring at a later period of life, would be termed pathological.'
  12. 'Handbuch der pathologischen Anatomie,' 1812, Vol. I., p. 9.
  13. Claude Bernard, 'Leçons sur la chaleur animale' (19th Lesson), 1875.
  14. To those who may wish to gain an attractive insight into modern conceptions of pathology—according to which disease is a relative term and its study a branch of biology—I would recommend Professor Woods Hutchinson's highly suggestive 'Studies in Human and Comparative Pathology' (1901).
  15. It is perhaps significant that while I had dealt with the 'organic variational tendency' Mr. Pearson prefers the vaguer term 'variability,' which enables him to bring in such matters as strength of pull and squeezes of hand, which are merely due to functional exercise.
  16. For the exact proportion of male to female still-born children in most civilized countries, see, e. g., Ploss, 'Das Weib,' 7th edition, 1901, Vol. I., p. 336.
  17. Joseph Clarke, 'Observations on some Causes of the Excess of the Mortality of Males above that of Females,' Philosophical Transactions, 1786. It may be said here that the very first attempt to weigh and measure infants accurately had only been made not so many years previously, by Roederer, in 1753.
  18. It seems unnecessary to deal with this point more in detail, not only because of the lack of sufficient data but because the establishment of this point is not necessary for the criticism of Professor Pearson's position. I expect to return to the point elsewhere, and hope that others, who may be more fortunate than I am in obtaining extensive data, will be able to deal with it on the lines I suggest.
  19. It must also be said that (as in the case of psychic genius) it is among the well-to-do classes that these very large infants are most usually found, not only because the parents tend to be larger among these classes, but because, as has lately been shown, other things being equal, women who rest during pregnancy tend to have larger children.
  20. In children dying at or soon after birth, as a result of undue pressure, hemorrhages or congestions are nearly always found in the internal organs, but they are not of necessity immediately fatal.
  21. It is somewhat unusual, Professor Pearson has remarked in a recent controversial paper ('Biometrika,' April, 1902, p. 323), 'in a discussion to give entirely different meanings to the terms originally used, and leaves your adversary to find out with what significance you may be using them.' It seems to occur sometimes however.