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

as the facts of Meristic Variation form the substance ..f i hi- \.-lume. it is unnecessary in this place to do more than refer t.. tli.' manner in which they exhibit the phenomenon of Dis- '•"iitiuuitv. < iin- or two instances must suffice to give some sug- • ii ..t" tin- -uliject, detailed consideration being reserved. Tartrepeated meristically form commonly a series, which is • •itln-r radial <>r linear. or disposed in some other figure derived fr.iin "i- compounded <>f these. For the purpose of this preliminary

treatment an instance "t I >i-cnnt inuous Variation in each of these ma !>.• taken. \ariations in the number of petaU of actinomorphic flowers exhiliit the I )i>c..ntintiity of Meristic Variation in perhaps its simplest form. Phenomena of precisely similar nature will hereafter be de- s.-nlied in animals. lmt Mich variati<»ns in flowers are so common and -. . a'-ce>sil»le that j-efereiice tu them mav with profit be made. In Fig. 7 such an example i~~ shewn.

It repi-e-.-nt- ;l Tulip haxin^ the parts of the flower formed in multiples .if tour, in-tead »f in multiples of three as normally. \ anati'in of this kind may In- seen in any field or hedgerow1. M-ri-tic Variation is here pre>eiite.| in its gre at est simplicity. Such a rase may well ser\e to illustrate some of the phenomena of D •• -lit iniiity. full lii'Tatmr mid lists df c;isrs si-f cspeciiilly MASTEUS, Verit'taMe '. \. l'.ili/]iliiilfii. It is iicrlnips nun. T.. -siiiy to refer to the fact that

fr.-iM mi ii--iiiiii.ti.>n l.y the members -f ...... series or whorl of the form and A form with four segments occurs as the offspring of a form with three segments. Such a Variation, then, is discontinuous FIG. 7. Diagram of the flower of a Tulip having all the parts in -4. because a new character, that of division into four, has appeai'ed in the offspring though it was not present in the parent. This new character is a definite one, not less definite indeed than that of division into three. It has come into the strain at one step of Descent. Instances in which there is actual evidence of such descent are rare, but there can be no question that these changes do commonly occur in a single generation, and, indeed, in many plants, as for example Lysimachia (especially L. nemorum), flowers having all the parts in -4 or in -6 may be frequently seen on plants which bear likewise normal flowers with the parts in -5.

Now such a variation as this of the Tulip illustrates a phenomenon which in the Study of Variation will often be met. We have said that the variation is discontinuous, meaning thereby that the change is a large and decided one, but it is more than this ; it is not only large, it is complete. The resulting form possesses the character of division into four no less completely and perfectly than its parent possessed the character of division into three. The change from three to four is thus perfected : from the form with perfect division into three is sprung a form with perfect division into four. This is a case of a total or perfect Variation.

This conception of the totality or perfection of Variation is one which in the course of the study will assume great importance, and it may be best considered in the simple case of numerical and Meristic Variation before approaching the more complex question of the nature of totality or perfection in Substantive Variation. The fact that a variation is perfect at once leads to the question as to what it might be if imperfect. Between the form in -3 and tinform in -4 are intermediates possible? and it' possible, do th.-y exist * Now by choosing suitable species of regular flowers, indiudual flowers may no doubt be found in which there are three l:ir. nents and one small oue ; or two normal segments and a

third dmded into two, making four in all. Such flowers are firstly \\hile cases of perfect transformation are common. But be- r rarity there is, further, a grave doubt whether they are in anv true sense intermediate between the perfect form in -:> and tin- |.".-rf. ct form in -4. After this again it must be asked whether 10 th. v do as a matter of fact occur as intercalated steps in the descent of the form in -4 from the form in -3 ? To the last questi..n a general negative may at once be given ; for though there is abundant evidence that Meristic Variations of many kinds arid in

ial degrees of completeness may be seen in the offspring of the aame parent, yel any one member of such a family group may -li.-w a particular Variation in its perfection, and the occurrence of anv int. im.-diate in the line of Descent is by no means necessary for ill.- production of the perfect Variation. 'I'., an-u. r the former question, whether or no forms imperfectly ividrd int.. four parts are in reality intermediate between those in -:{ and those in -4, a knowledge of the mechanics of the I in ..-.--. of 1 )i\ i-ion is required. Such knowledge is as yet entirely \\antiir_r. and di-'-ussion of this matter must therefore be prematun-. With niu.-h hesitation I have decided to make certain i. tl. xi'.n- .'ii tlh- -ulijert, which will be found in an Appendix to tlii- w.irk. Tli-- may perhaps have a value as suggestions to oth.-r-. though from th.-ir theoretical nature they can find no place

Th. !•• i- how.-ver an. it h< -r class < if cases which are intermediate in a ditf«-n-nt wav. In tin- Tulip described above the quality of divi-i-.n int.. I- WBS piv>.-m in all tinfloral organs. This is not alwavs th. case, tor a M.-ristic Variation may be present in one as, though it is aljsent in -ome or all of the others, and this i- a phenomenon fn-.|ii.-nt ly i-fi-urring. Nevertheless, tli..tiL,di onl\ partially distrilmted, a Variation inav still be dis- |ila\i-d in it-> totality in the parts wherein it is present. The pan- "t a -in-^liuliorl, the <-al\\ for .-\ainple, may undergo a coiii|.l. t.- N'ariatioii, while the corolla and other parts are unehang.-d. In the same \\ay single memlu-rs of a radial series, as a ).. tal for example, may undergo a complete Variation while the oth'-r in. mil. i- of tinseries are unchanged. The same will be -h.'un hen-alt, r t» be true of animals also.

K.'i iii-iaiic.-, the normal number of the parts in the disc of Am-i'liii is four, but the whole liodv ma\ \»- divided instead into -i\ .T some other number of pans. Examples are also found in uliich tinparts (,f one-half or of one quadrant are arranged in the new numli.-r, whd«- the remainder is normal; and, as in flowers, this new number may prevail in some or in all of those systems of organs which are disposed around the common centre. Before speaking further of the totality or perfection of Variation it will be well to give an illustration of Discontinuous Meristic Variation as it occurs in the case of a linear series of parts. As such an illustration the case of the variation in the number of joints in the tarsus of the Cockroach (Blatta) may be taken. This variation has been the subject of very full investigation by Mi- ll. H. Brinclley. The tarsus of the Cockroach is normally divided into five joints, but in about 25 per cent, of B. americana (and in a smaller proportion of several other species) the tarsus of one or more legs is divided into only four joints, though the total length may be the same as that of the corresponding leg of the other side, Fig. 8. Between the five-jointed form of tarsus and the four-jointed form no single case in any way intermediate was seen. The whole

FIG. 8. Tarsi of the third pair of legs in a specimen of Blatta americana, I. the left tarsus, having the normal, or 5-jointed form ; II. the right tarsus, having the 4-jointed form. evidence will be given in full in the proper place and raises many questions of great interest ; but that which is important to our present consideration is the fact that the Variation is here undoubtedly discontinuous, arising suddenly as a total or perfect Variation, from the five-jointed form to the four-jointed. Here the variation, though total as regards the limb in which it is present, is not total as regards all the legs taken together. For commonly only a single leg had a four-jointed tarsus, and only one specimen was met with in which all six legs thus varied, and one specimen only shewed the variation in five legs.

In speaking of such a Variation as a perfect Variation several things are meant. First, it is meant that the tarsus of the new pattern is as distinctly divided into four joints as the normal is into five. In addition to this the statement that the varying limb is perfect conveys a number of ideas that cannot be readily formulated ; for mple. that tinjoints are to all appearance properly proportioned and -' -i \ ic. -able, shewing no sign of untitness: they have in fact nnii-li tinsiime appearance as they have in those of the Orthoptera in \\hich tintarsus is normally four-jointed. But besides these attributes, which though useful enough for ordinary description are -till in their nature funnies, and of no precise application, th- • mother which in the case of these varying legs we are

.-lit it l.-d to make. We have said that these four-jointed tarsi are ill a]i]n arance normal, .save for the number of the joints. Now the niea-iireinents which, at my suggestion, Mr Brindley has been kind enough to make, entitle us to go beyond this, and to assert that tintour-jointed tarsus has another character by reason of \\hich it i- actually in a sense a "normal" form. A brief consideration of tinwill clearly illustrate the meaning of the term "perfeeiion ' ;iji|,lj,-il to Variation.

W'- -aw above that in a monomorphic form, the frequency with which, in respect of any given character, it departs from its mean condition follows a curve of Frequency of Error. This is, indeed, \\hat i- meant by the statement that the mean condition is a normal. Taking the tive-jointed tarsus, measurements shewed that the ratio of the length of any given joint to the length of the whole tarsus varied in this way about a mean value. Measurement of the jointof the four-jointed form shewed that the ratios which they bear to the total length of their respective tarsi vary in a similar way about their mean values, and that there is thus a "normal" four-jointed conditi< n just as there is a "normal" fivejointed condition. In the .-aiue way, then, that the ratio of the length of e.,e|i of the ti\e joints to that of the whole tarsus is not alua\s identical but exhibits small variations, so the ratios of the

ral joints of the four-jointed tareua to the length of the whole tarsus al-o \ar\, but in each case the ratio has a mean value which is approached with a frequency conforming to a curve of Ki .' i <r. The measurements established also another fact which is of consequence to an appreciation of the nature of totality in Variation. It not only appealed that the departures from the mean \alinof these ratioin the tour-jointed variety were distributed about the mean in the same \'vav as those of the fivejoint. -d form, but it was also shewn that the absolute varia- lOnfl tioni the mean \abies of these ratios were not on the "'•• greater in the tour-jointed tarsi than in the five-jointed In other words, the tour-jointed tarsus occurring thus sporadically, as a variety, is not |e>s definitely constituted than ive jointed type, and the proportions «,f its several joints oatant. It i- scarcelj accessary to point out that

these facts give no support to the view that the exactness or perfection with which the proportions of the normal form are approached is a consequence of Selection. It appears rather, that there are two possible conditions, the one with five joints and the other with four, either being a position of Organic Stability. Into either of these the tarsus may fall; and though it is still conceivable that the final choice between these two may have been made by Selection, yet it cannot be supposed that the accuracy and completeness with which either condition is assumed is the work of Selection, for the "sport" is as definite as the normal.

This interesting case of Meristic Variation in the tarsus of the Cockroach illustrates in a striking way the principle which is perhaps the chief of those to which the Study of Variation at the outset introduces us. We are presented with the phenomenon of an organ existing in two very different states, between which no intermediate has been seen. Each of these states is definite and in a sense perfect and complete ; for the oscillations of the fourjointed form around its mean condition are not more erratic than those of the normal form. Now when it is remembered that just such a four-jointed condition of the tarsus is known as a normal character of many insects and especially of some Orthoptera, it is, I think, difficult to avoid the conclusion that if the four-jointed groups are descended from the five-jointed, the Variation by which this condition arose in them was of the same nature as that seen as an individual Variation in Blatta ; that as the modern phenomenon of the individual Variation which we see, so that past phenomenon of the birth of a four-jointed race, was definite and complete, and that the change whose history is gone, like the change to be seen to-day, was no gradual process, but a Discontinuous and total Variation1.

1 Since this Section was written it has seemed possible that the account given above may be found to need an important modification. It is well known that BUittd, in common with many other Orthoptera, has the power of reproducing the antennae and legs after amputation or injury, and we have made some observations shewing that the tarsi of these regenerated legs sometimes, if not always, contain four joints. The question therefore arises whether the 4-jointed tarsus is a truly congenital variation, and not rather a variation introduced in the process of regeneration, somewhat after the manner of a budvariation. To determine this point a considerable number of immature specimens were examined, and it was found that the percentage of individuals with 4-jointed tarsi is considerably less in the young than in the adult. These facts lend support to the view that the 4-jointed condition is not congenital. A quantity of individuals were also hatched from the egg-cocoons, and among them there has thus far been found no case of 4-joiuted tarsus. On the other hand the total number thus hatched is not yet sufficient to create any strong probability that none are ever hatched in the 4-joiuted state. We have also seen the 4-joiuted tarsus in three very young individuals, which, to judge from their total length, must have been newly hatched. The statistics shew besides that the abnormality is distinctly commoner in females than in males, and that it is commoner in the legs of the 2nd pair than in the 1st, and much more common in the 3rd pair of legs than in the 2nd. These facts somewhat favour the view that the variation may be congenital. It seems also exceedingly improbable that in the specimen with all the tarsi 4-joiuted, the six legs could each have been lost and renewed. There seems on the whole to be a pre-

The application of the term Discontinuity to Variation must n.,t In- misunderstood. It is not intended to affirm that in dis- •inuou- Variation thenran l»f between the variety and the type no intermediate form, or that none has been known to occur. ;md it is not even necessary for the establishment of Discontinuity ih.it the intermediate forms should be rare relatively to the !„ it., t form of the variety, though in cases of discontinuous Variation this is generally the case: but it is rather meant that tinperfect form of the variety m<iy appear at one integral step in Descent, either without the occurrence of intermediate gradations, "i .it least without the intercalation of such graduated forms in th,' pedigree.

In the case of the tarsus of Blntta we have seen an example ..t a i,.tal and complete Variation affecting single members of a series of repeated parts, not collectively, but one or more at a tun'-1. Such an instance of a Meristic Variation occurring in a -tat,' which is total as regards members of a series but not total a- regards the whole series finds many parallels among Sub-taut i\>- Variations, as, for example, that of the Crab (Cancer ;//•</.-) bearing the right third maxillipede fashioned as a chela, whit.- tli.- left third maxillipede was normal. Variations of this nature in plants are of course well known to all.

V a previoiiplace (Section vn.) allusion was made to the familiar but very curious analogy between members of a series of M eristic parts and separate organisms. The facts of Variation bring oin thianalog in many -insular ways, and in speaking of the totalitv of Variation it is necessary to hear these facts in nn'nd. N •• only anthere abundant instances of independent division or multiplication ,.f -ingle m< -miters of Meristic series, but as ha- l>e. -n -aid. -ingle memberof such series may thus independently an, 1 singly undergo qualitative or Substantive Variation, being treated in tinph\-ical sy-tem of the body as though they - -pai-ate units. In Variation, therefore, though it will be

• ption tliut the variation may .-it !ea-t smn, .times be congenital. Supposing howe\er Unit tliis -hull lie found hereafter not to be the ease, I do not think that tinillin, -t. Mililniwn fi-,,m the t'ui't-i \vill he less vuliil. The conclusions as to the ileliniteness of the two t\|n^, ;ill>l the Vel :| t ion -h i ] is of the several parts of each il purts of the other, would still hold ^ood. Thenare besides in other I'"' -leeof similar nuinei ieai Variation, as for example, in the number of

jointin the antenna, of I'rioiiida', where the hypothesis of change on renewal is imp., -sih],.. from which a -imiliir argument ini^ht be drawn; but on the whole I '"^ • i"' as it stands, taking the case of Blntta as an •-ily a, ,-.•— ilile and be.-au-e. from die fewness of the joints roin-i-ini d. tli,- i-siirs are singularly char. found that members of Meristic series may vary simultaneously and collectively — and this is one of the most important generalizations which result from the Study of Variation — yet it is also true that in Variation single members of such series may vary independently and behave as though they possessed an " individuality" of their own. If ever it shall be possible t<> form a conception of the physical processes at work in the division and reproduction of organisms, account must be taken of both of these phenomena.

I know no way in which the nature of Discontinuity in Variation and the position of intermediate forms may be so well illustrated as by the closely parallel phenomenon of Sex. In the case of Sex in the higher animals we are familiar with the existence of a race whose members are at least dimorphic, being formed either upon one plan or upon the other, the two plans being in ordinary experience alternative and mutually exclusive. Between these two types, male and female, there are nevertheless found intermediate forms, "hermaphrodites," occurring in the higher animals at least, as great rarities. Now though these intermediate forms perhaps exist in gradations sufficiently fine to supply all the steps between male and female, it cannot be supposed that the one sex has been derived from the other, and still less that the various stages of hermaphroditism have been passed through in such Descent. Besides this, even though there is an accurate correspondence or homology between the several organs which are modified upon the one plan in the male and upon another in the female, and though this homology is such as to suggest, were we comparing two species, that the one had been formed from the other, part by part, yet by the nature of the case such a view is here inadmissible : for firstly it is impossible to suppose that either sex has at any time had the organs of the other in their completeness, and secondly it is clear that any hypothetical common form, by modification of which both may have arisen, must have been indefinitely remote and could certainly not have possessed secondary sexual organs bearing any resemblance to those now seen in the higher forms. All this has often been put, but the application of it to Variation is of considerable value.

For in the case of Sex there is an instance of the existence of two normals and of many forms intermediate between them, occurring in a way which precludes the supposition that the intermediates represent stages that have ever occurred in the history of the two forms. In yet another way Sex supplies a parallel to Variation. As we know, the sexes are discontinuous and occur commonly in their total or perfect forms. Now just as members of a Meristic series may present total variations independently of each other, so may single members of such a series present opposite secondary sexual characters, which may nevertheless be in each case complete.

The best known instance of this that of gynandromorphic in-e.-ts, in \\hi.-h the characters of the whole or part of one side ..i the l""i\, \vm:,'> and antenna-, are male, while those of the ,,t! e are female. 1 1« •nmrkable instances of a similar phenomenon have been recorded among bees and will be described later. A- i- u. 11 known, the organs and especially the legs of the sex- f. male* «»r workers are formed differently from those of the dr.. ii.--. but there are cases of individuals having some of the j.ari- and appendages formed on the one plan and some on the other. Thus in these individuals, which are in a sense intermediate between workers and drones, the characters of the two mav still be not completely blended, the male type pre- \ailiii:,' in .some parts, and the female in others. In the Discontiniiit v of Substantive Variation will be found examples of imperfect blending of variety and type closely comparable with tin- <-ase of the imperfect blending of Sex.

Si GOES! K>NS AS TO THK NATUHE OF DISCONTINUITY IX VARIATION. 'I'lnob-.-nations at the end of Section xi, regarding the Dis- .-..iitinuity »f M eristic Variation lead naturally to certain reflexions .1- i.. i In- nature of Discontinuous Variation in general. In the <>t' the < 'orkroaeh tarsus, there given, it appeared that just as the -tnieture of the typical form varies about its mean condition, BO the structure of the \arietv \ ai'ies about another mean condition. Tliifa'-t. \\hieh in tingi\en instance of Meristic Variation is so dear, at oiicc suggests an in.|iiiry whether this is not the usual coin-.- of Discontinuous Variation, and, indeed, whether Discontinuity in Variation docs not mean just this, that in varying the

inism passes from a form \\hich is the normal for the type to another form \\lii.-h i- a normal for the \ariety. Such transitions plainly occur in many cases of McHst ic Variation, and in a considerable number of Sub-taut i\e Variations there will be found to be indicationthat the phenomenon is similar. It is true that at the pn - nt stage of the in.|iiiry the .\ idence has the \alue rather of 'i"ii than of proof, but the suggestion is still very decided and it i- scarcely po»ihle to exaggerate the importance of even tinslender clue.

In stating the problem of Species at the ben-inning of this inquiry it \\a- -aid that the forms of living things, as we know them, constitute a di-continuou> series, and it is with the origin "t tin- Discontinuity of the series that the solution of the main pioKlrin is lar-.-K concerned. N.,\v tinevidence of Discontinuous \anation auggeets that organisms may vary abruptly from the definite form of the type to a form of variety which has also in some measure the character of definiteness. Is it not then possible that the Discontinuity of Species may be a consequence and expression of the Discontinuity of Variation ? To declare at the present time that this is so would be wholly premature, but the suggestion that it is so is strong, and as a possible light on the whole subject should certainly be considered.

In view of such a possible solution of one of the chief parts of the problem of Species it will be well to point out a line of inquiry which must in that event be pursued. If it can be shewn that the Discontinuity of Species depends on the Discontinuity of Variation, we shall then have to consider the causes of the Discontinuity of Variation. Upon the received hypothesis it is supposed that Variation is continuous and that the Discontinuity of Species results from the operation of Selection. For reasons given above (pp. 15 and 10) there is an almost fatal objection in the way of this belief, and it cannot be supposed both that all Variation is continuous and also that the Discontinuity of Species is the result of Selection. With evidence of the Discontinuity of Variation this difficulty would be removed.