Materials for the Study of Variation Treated with Especial Regard to Discontinuity in the Origin of Species

It will be noted also that it is manifestly impossible to suppose that the perfection of a variety, discontinuously and suddenly occurring, is the result of Selection. No doubt it is conceivable that a race of Tulips having their floral parts in multiples of four might be raised by Selection from a specimen having this character, but it is not possible that the perfection of the nascent variety can have been gradually built up by Selection, for it is, in its very beginning, perfect and symmetrical. And if it may be seen thus clearly that the perfection and Symmetry of a variety is not the work of Selection, this fact raises a serious doubt that perhaps the similar perfection and Symmetry of the type did not owe its origin to Selection either. This consideration of course touches only the part that Selection may have played in the first building up of the type and does not affect the view that the perpetuation of the type once constituted, may have been achieved by Selection.

But if the perfection and definiteness of the type is not due to Selection but to the physical limitations under which Variation proceeds, we shall hope hereafter to gain some insight into the nature of these limitations, though in the present state of zoological study the prospect of such progress is small. In the observations which follow I am conscious that the bounds of profitable speculation are perhaps exceeded, and I am aware that to many this may seem matter for blame ; but there is, in my judgment, a plausibility in the views put forward, sufficient at least to entitle them to examination. They are put forward in no sense as a formulated theory, but simply as a suggestion for work. It is, besides, only in foreseeing some of the extraordinary possibilities

that lie ahead in the Study of Variation, that the great value of tinin. -th. ..| can be understood. h ha- 1..-OI s.-en that variations may be either Meristic or Sub-taut i\. •. and that in each group discontinuous and definite \\ are imw seeking the factors which determine this totality and .1. tin.- th.- forms assumed in Variation. In this attempt we may, l.v jirl.it tariK confining our first notice to very simple cases, recog- n"i/e at 1. a-t two distinct factors which may possibly be concerned

ii, th;- : initiation. Of these the first relates to Meristic I. Possible nature of the Discontinuity of Meristic Variation. Looking at simple cases of Meristic Variation, such as that of tin- Tulip or of Aurelia, or of the Cockroach tarsus, there is, I think, a fair sugg( -ii"!» that the definiteness of these variations is determini-'l mechanically , and th&t the patterns into which the tissues of animalandivided represent positions in which the forces that . tf.-.-t tin di\i.-ion are in equilibrium. On this view, the lines or plan-- of di\ision would be regarded as lines or planes at right angl. •> t.. tindirections of the dividing forces; and in the lines of M. ii-tii- ]>i\i-ion we are perhaps actually presented with a map of the lines of tho-e forces of attraction and repulsion which di-tcrmimth>' number and positions of the repeated parts, and ti-.in ulii.-h Symmetry re.-ults. Jf the Symmetry of a living bod}' \\t-n- tlni- i. '•"•_; ni/. d as of the same nature as that of any symni. nical -\-teiii «'f mechanical forces, the definiteness of the sym-

in. n\ in Meri-tic Variation would call for no special remark, and the \» Tlei-tion of th>- -ymmetry of a Tulip \\ith its parts divided int.. four, though occurring suddenly as a " sport," would be recogni/. d a> in nowis.- more sin--nlar than the symmetry of the type. I loth alike would tln-n In seen to owe their perfection to medianical condition- and not to Selection or to any other gradual process. It reason for adopting such a view of the physics of l>i\i-ion >hould a|i)iear. the frei|iiency with which in any given l-'im a particular pattern of Division or of Symmetry recurs, \\oiild lie toiind to l.e determined by and to be a measure of the stability of the forces of l)i\ivj,,n \\hen disposed in that particular patt.rn. It will ..f course lie understood that in these" remarks H" !!"ii i- offered as t" the causes \\hich determine whether

a tis-.ii.- shall di\id.- into four or into three, but merely as to the Conditions "f pert'. «-t ion ..f the di\ ision iii eit her case. It will also 1" clear that though the symmetry of a flower or of any other tissue depend- a I- -\ minelrical gr.. \\tli, it is primarily dependent on tin- -\miiietn of its primary di\isions. ujion which symmetrical ;ji..uili and -econdary sy nniiet rical di\i-ions follow. It would be interesting and I believe profitable to examine somewhat further the curiously close analogy between the symmetry of bodily Division and that of certain mechanical systems by which close imitations both of linear and of radial segmentation can be produced ; and though to some this might seem overdaring, the possibility that the mechanics of bodily Division are in their visible form of an unsuspected simplicity is so far-reaching that it would be well to use any means which may lead others to explore it.

And even if at last this suggestion shall be found to have in it no other element of truth, it would still be of use as a forcible presentation of the fact, which when realized can hardly be doubted, that among the factors which combine to form a living body, the forces of Division may be distinguished as in their manifestations separable from the rest and forming a definite group. For, already (Section v.) it has been pointed out that the patterns of Division or Merism may be changed, while the Substance of the tissues presents to our senses no difference. The recognition of this essential distinctness of the Meristic forces will, I believe, be found to supply the base from which the mechanics of growth will hereafter be attacked.

The problems of Morphology will thus determine themselves into problems in the physiology of Division, which must be recognized together with Nutrition, Respiration and Metabolism, as a fundamental property of living protoplasm. To sum up : there is a possibility that Meristic Division may be a strictly mechanical phenomenon, and that the perfection and Symmetry of the process, whether in type or in variety, may be an expression of the fact that the forms of the type or of the variety represent positions in which the forces of Division are in a condition of Mechanical Stability.

2. Possible nature of the Discontinuity of Substantive Variation. Passing from the phenomena of Division and arrangement to those of constitution or substance we are, as has been said, again presented with the phenomenon of discontinuous or total Variation, and we must seek for causes which may perhaps govern and limit this totality, and in obedience to which the Variation is thus definite. Now as in the case of Meristic Variation, by arbitrarily limiting the examination to those cases which seem the simplest it appears that there is at least an analogy between them and certain mechanical phenomena, so by similarly restricting ourselves to very simple cases there will be seen to be a similar analogy between the discontinuity of some Substantive Variations and that of chemical discontinuity. It is on the whole not unreasonable to expect that the definiteness of at least some Substantive Variations depends ultimately on the discontinuity of chemical affinities. To take but one instance,

that of colour, we are familiar with the fact that the colours of many organic substances und.-rg.. definite changes when chemi- 1 on by reagents, and it is not suggested that the d. Unite,,, .ss and discontinuity of the various colours assumed is dependent on anything but the defiriiteness of the chemical changes undergone. The changes of litmus and many vegetable blues to red on treatment with acids, of many vegetable yellows i.rown on treatment with alkalies, the colours of the series of Udies produced by the progressive oxidation of bihverdm are tamiliar examples of such definite colour-variations.

• I on us that the definiteness of colour-variation is a consequence of the defiuiteness of the chemical changes undergone. \ one doubts that the orange colouring matter of the variety the Iceland Poppy (P. midicaule) is a chemical derivative from the yellow colouring matter of the type. It is not questioned that in such cases a definite alteration in the chemical conditions in which the pigment is produced determines whether tli.- llow.T shall be orange or yellow; and I think it is reasonable to i that the frequency with which the flowers are either

;..\\ or orange as compared with the rarity of the intermediate >|,;. ,11 expression of the fact that the yellow and orange forms bhe colouring matter have a greater chemical stability than the int. iin.ili.it> forms of the pigment, or than a mixture of the two If then it should happen, as we may fairly suppose it illicit, that the orange form were to be selected and established as it would owe the definiteness of its orange colour and tin- ]>ivri>i.>n of its tint, not to the precision with which Selection had eh"s.-n this part ienlar tint, but to the chemical discontinuity of \\liidi th. . .li-inally di-i-oiit innous Variation was the expression.

To |':i~-- troin ili.- case of a sport to that of Species, it is well known that of the many S. African butterflies of the genus /.' * J ntlinclni /•/>-, < )range-tips). some have the apices or tips of th.- foiv -\\in-s orange-red (tor example, E. danae), while in othe^ th'-y are purple (tor example, /•,'. lone}. Upon the -s that tintransition from orange to purple, or vice versa, had li.-. n e.nit innoiisly effected liy 1 he successive Selection of niinutf \ariatioii-. \\.- arc met by all tin.- difficulties we know so

\\i-ll. \Vhy is pnrplr a - I colmir for this creature ? If purple i-~ a ^on.l c..lonr and n-d i> a -oml culmir. ho\v did it happen that at -omr tini.' or other all tinintermediate shades were also good .•ii-H-h to ha\e been selected? and so on. These and all the grnate ilith'cnlti.- are o|»-n.'d up at once, and though they have l»'ii met iii the fa-hion we know, they have scarcely been overcome. Hut at tinoutset thU view assumes that every interini-diate may e\i>t and has existed, an assumption which is

ituit'.iis and hardly in armrdanre with the known fact that eh. miral \<i<- at.- 1 1 , •, pie] , i ] \ disron t inuous. When besides this it is known that Variation may be discontinuous, I submit that it is easier to suppose that the change from red to purple was from the first complete, and that the choice offered to Selection was between red and purple ; and that the tints of the purple and of the red were determined by the chemical properties of the body to which the colour is due. This case is a particularly interesting one in the light of the fact that, as Mr F. G. Hopkins has lately shewn me, this purple colour, dissolved in hot water, leaves on evaporation a substance which gives the murexide reaction and cannot as yet be distinguished from the substance similarly derived from the orange or yellow colouring matters of Pieridce in general. As was stated above, Mr Hopkins has shewn that these yellows are acids, allied to mycomelic acid, a derivative of uric acid, and therefore of the nature of excretory products. Whether the purple body is related to the yellow or to the orange as a salt is to an acid, or otherwise, cannot yet be affirmed ; but if the difference between them is a chemical difference, which can hardly be doubted, there is at least a presumption that the discontinuity of these colours in the several species, is an expression of the discontinuity of the chemical properties of this body. The possibility that from such bodies a series of substances might perhaps by suitable means be prepared in such a way as to represent many or even all intermediate shades, does not greatly affect the suggestion made ; for even in such series it is almost certain that points of comparative stability would occur, and Discontinuity would be thus introduced.

The case of Colour has been taken in illustration because it is the simplest and most intelligible example of the possibility that the Discontinuity of some Substantive Variations is determined by the Discontinuity of the chemical processes by which the structures are produced. It is true that perhaps no species has been rightly differentiated by colour alone, but colour is still one of the many characters which go to the distinguishing of a species, and it is precisely one of the characters whose significance and delimitation by Natural Selection is most obscure. Moreover by the fact that in the case of these yellow and red Pieridse the colours are of an excretory nature, we are reminded that Variation in colour may be an index of serious changes in the chemical economy of the body, and that when an animal is said to be selected because it is red or because it is purple, the real source of its superiority may be not its red colour or its purple colour, but other bodily conditions of which these colours are merely symptoms. By those who have attempted to reconcile the phenomena of Colour with the hypothesis of Natural Selection this fact is too often overlooked.

But though it may reasonably be supposed that much of the Discontinuity of Variation and some of the Discontinuity of Species arise through discontinuous transition from one state of iii.-.-l,:li,i,-;il or chemical stability to another state of stability, there M.-V, -rtheless remain large classes of discontinuous variations, and - .-it'ii.' Differences still more, whose Discontinuity bears no Malogy with these. To these phenomena inorganic Nature

ra IK. parallel. We may see that they are discontinuous and tint tin-itcourse is in some way controlled, but as to the nature of this control we can make no guess. Though the resemblance may be misleading, it is nevertheless true that in living Nature there are other phenomena, those which present a Discontinuity closely comparable with tint of nianv variations. In problems of disease we meet again tin- >ame problem which we meet in Variation, namely, changes whi.-h may be complete or specific, though occurring so suddenly \clmle the hypothesis that Selection has been the limiting cause. All this is" familiar to everyone who has considered the |n-< .Kit -in of Species.

For though, like discontinuous variations, the manifestations ..t' -p--eiiic disease are not always identical, but differ in intensity ami d' arying about a normal form, still these manifestations inav b.- -p--citic iu the sense in which the term is used with reference to the rharai-t'Ts of Species. If we exclude those diseases whose -p'-ejtie characters are now known to be the result of the invasion . .t' -p.-citic or^ani-ms, there still remain very many which are known and rece,_rni/ed l»v definite and specific symptoms produced in the l>"dv. though thne is as yet no evidence that they are due to specific organisms. [Of cour.-e it' it were shewn that these diseases also result from tinaction of specific organisms, they then only pre-.-nt to H- aijain tinoriginal problem of Species ; for if the iie-- or Specie-, of a disease is due to the definiteness, or . ,,f th.- micro-organism which causes it, the cause of that of the 1 1 1 ji To-or^ai iisi ii remains to be sought, and we are -imply left with a particular case of the general problem of Specie-.) Mut in the meantiuir we can see that the manifestations specific : ami while \v do not know that they result from causes themselves -pe-'jtic, the nature of the control in obedience to which tip -pecitic is unknown.

The parallel between disease and Variation may be misleading, l.ut this much at leasl may fairly be learned from it: that the -y-t,-m of an organixed being is such that the result of its disturbance may lie specific. And in the end it may well be that the prolilem of Species will be solved by the study of pathology: tor the likene-,-, between Variation and disease goes far to Mipptu-t th.- view which Yirchow has forcibly expressed, that " every deviation from the type of the parent animal must have it-, foundation on a path .logical accident1."

Enough has now been said to explain the aim of the Stud\ of Variation, and to shew the propriety of the choice of the facts of Meristic Variation as a point of departure for that study. Before leaving this preliminary consideration, reference to some cognate subjects must be made. It has been shewn that in view of the facts of Variation, some conceptions of modern Morphology must be modified, while others must be abandoned. With the recognition of the significance of the phenomena of Variation, other conceptions of biology will undergo like modifications. As to some of these a few words are now required, if only to explain methods adopted in this work.

It has been the custom of those who have treated the subject of Evolution to speak of " Heredity " and " Variation " as two antagonistic principles ; sometimes even they are spoken of as opposing "forces." With the Study of Variation, such a description of the processes of Descent will be given up, even as a manner of speaking. In what has gone before I have as far as possible avoided any use of the terms Heredity and Inheritance. These terms which have taken so firm a hold on science and on the popular fancy, have had a mischievous influence on the development of biological thought. They are of course metaphors from the descent of property, and were applied to organic Descent in a time when the nature of the process of reproduction was wholly misunderstood. This metaphor from the descent of property is inadequate chiefly for two reasons.

First, by emphasizing the fact that the organization of the offspring depends on material transmitted to it by its parents, the metaphor of Heredity, through an almost inevitable confusion of thought, suggests the idea that the actual body and constitution of the parent are thus in some way handed on. No one perhaps would now state the facts in this way, but something very like this material view of Descent was indeed actually developed into Darwin's Theory of Pangenesis. From this suggestion that the body of the parent is in some sort remodelled into that of the offspring, a whole series of errors are derived. Chief among these is the assumption that Variation must necessarily be a continuous process ; for with the body of the parent to start from, it is hard to conceive the occurrence of discontinuous change. Of the deadlock which has resulted from the attempt

to interpret Homology on this view of Heredity, I have already -poki-n in Section vi. - ••-.ndly, tinmetaphor of Heredity misrepresents the essential phi-iioiin-non of reproduction. In the light of modern investigation-, an. I especially those of Weismann on the continuity of the in -• -i-lls, it is likely that the relation of parent to offspring, it it h;us any analogy with the succession of property, is ratln-r thai of trustee than of testator.

II' reafter, perhaps, it may be found possible to replace this metaphor by some more correct expression, but for our pr.--.-nt purpose this is not yet necessary. In the first examination of the facts of Variation, I believe it is best to att'-nipt no particular consideration of the working of Heredity. Tipphenomena of Variation and the origin of a variety must ii'-r,-— ,-irily be studied first, while the question of the perpetuation of i lie variety properly forms a distinct subject. Whenever in tincases given, observations respecting inheritance are forth- "•••miiig they will be of course mentioned. But speaking of discoiii unions Variation in general, the recurrence of a variation in offspring, either in the original form or in some modification of it. has been seen in so many cases, that we snail not go far \\i-ong j,, at least assuming the possibility that it mat/ reappear in tinoffspring. At the present moment, indeed, to this statein. -nt th.'iv is little to add. So long as systematic experiments in bivi-ding are wanting, and so long as the attention of naturalists i- limit. -d to the study of normal forms, in this part of biology which is perhaps of greater theoretical and even practical im-

Around 'lii1 term Reversion a singular set of false ideas have gathered themselves. ()u the hypothesis that all perfection and completenes of i,,nu Or of eon-elation of parts is the work of •ti"ii it is difficult to explain the discontinuous occurrence tion is inadmissible is shewn at once by the frequent occurice U discontinuous Variation, of forms which though equally if this supposition is made, it cannot be made again for the equally perfect forms with three petals, and the rest1.

The hypothesis of Reversion to account for the Symmetry and perfection of modern or discontinuous Variation is made through a total misconception of the nature of Symmetry. There is a famous passage in the Descent of Man, in which Darwin argues that the phenomenon of double uterus, from its perfection, must necessarily be a Reversion. "In other and rarer cases, two distinct uterine cavities are formed, each having its proper orifice and passage. No such stage is passed through during the ordinary development of the embryo, and it is difficult to believe, though perhaps not impossible, that the two simple, minute, primitive tubes could know how (if such an expression may be used) to grow into two distinct uteri, each with a wellconstructed orifice and passage, and each furnished with numerous muscles, nerves, glands and vessels, if they had not formerly passed through a similar course of development, as in the case of existing marsupials. No one will pretend that so perfect a structure as the abnormal double uterus in woman could be the result of mere chance. But the principle of reversion, by which long-lost dormant structures are called back into existence, might serve as the guide for the full development of the organ, even after the lapse of an enormous interval of time-." fiescent of Man, vol. i. pp. 123 and 1'24.

This kind of reasoning has been used by others again and again. It is of course quite inadmissible ; for by identical reasoning from the perfect symmetry of double monsters, of the single eye of the Cyclopian monster, and so on, it might be shewn that Man is descended from a primitive double vertebrate, from a one-eyed Cyclops and the like. For other reasons it is likely enough that double uterus was a primitive form ; but the perfection and symmetry of the modern variation to this form is neither proof nor indication of such an origin. Such a belief arises from want of knowledge of the facts of Meristic Variation, and is founded on a wrong conception of the nature of symmetry and of the mechanics of Division. The study of Variation shews that it is a common occurrence for a part which stands in the middle line of a bilaterally symmetrical animal, to divide into two parts, each being an optical image of the other: and that conversely, parts which normally are double, standing as optical images of each other on either side of such a middle line may

1 For a full account of such facts, see a paper by Miss A. BATESON and myself On Variations in P'loral Symmetry. Journ. Linn. Soc., xxvui. p. 386. • This extraordinary passage is scarcely worthy of Darwin's penetration. If read in the original connexion it will seem strange that it should have been allowed to stand. For in a note to these reflexions on Reversion (Descent, i. p. 125) Darwin refers to and withdraws his previously expressed view that supernumerary digits and mamrnse were to be regarded as reversions. This view had been based on the perfection and symmetry with which these variations reproduce the structure of putative ancestors. It was withdrawn because Gegenbaur had shewn that polydactyle limbs often bear no resemblance to those of possible ancestors, and because extra mammae may not only occur symmetrically and in places where they are normal in other forms, but also in several quite anomalous situations. In the light of this knowledge it is strange that Darwin should have continued to regard the perfection and symmetry of a variation as evidence that it is a Reversion.

I- would probably help the science of Biology if the word 1;. . raion' and theideas which it denotes, were wholly dropped, uts until Variation has been studied much more fully than it has yet been. In the "light of what we now know of the process of reproduction the phrase is almost meaningless. We suppose that a tin stock gives off a number of individuals which vary about a normal; and that after having given them off, it begins to pve ..IV individuals varying about another normal. We want