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

!>• di-taili-d ln-iv, and it will therefore be enough to give an iH- of ii> chief fi'atuivs and to point out the bearing of this • •la— - of rvid.'iirimi tlii- -ul)jf<-t of Meristic Variation in general. INDIVIDUAL Variation in the fundamental number of members constituting a Linear Series of segments can only be recognized in those forms which at some definite stage in their existence cease to add to the number of the series. Hence in a large proportion of the more fully segmented invertebrates this phenomenon cannot be studied, for in many of these, as for instance in Chilognatha, and in most of the Chsetopoda the formation of new segments is not known to cease at any period of life, but seems to continue indefinitely. On the other hand, while in Insecta, and in Crustacea excepting the Phyllopods, the fundamental numbers are definite, no case of individual Variation in them has been observed.

Between these two extremes, there are animals in certain classes, for example, Peripatus, some of the Chilopoda among Myriapods, Aphroditidse among Annelids, and some of the Branchiopoda among Crustacea, in which the number of segments does not increase indefinitely during life, but is nevertheless not so immutable as in the Insects and the majority of Crustacea. In the forms mentioned, certain numbers of segments, though not the same for the whole family, are characteristic of certain genera, as in the case of the Chilopoda (excepting Geophilidre), or of certain species, as in some of the Peripati. But besides this, in some of the forms named, e.g., the Geophili and Peripatus edwardsii, individual Variation has been recorded among members of the same species. It is unfortunate that for many of the forms in which Variation of this kind possibly takes place, no sufficient observation on the point has been made, but as examples of a phenomenon which, on any hypothesis, must have played a chief part in the evolution of these animals, the few available instances are of interest.

*1. Peripatus. The number of segments which have claw-bearing ambulatory legs differs in different species of this genus. While, inoi, o\, T, in some of the species the number appears to be 'very con-taut tintin- -peejes, in the case of others, great individual \ai iation i- -een to oerur. SEDOWICK*8 observations in the caseof /' 1,1,1;! /-'1 1 1 -he\\ conclusively that these variations cannot be a- ribed t,. difference in age. There is besides no ground for suppo-ing that iii'Tease in the number of legs occurs in any species after birth, and it i>- in fact practically certain that this is not the case. In Pi'n'jHitus cupensis, which was exhaustively studied by S. cl_-\\ i.-k, the appendages arisein the embryo successively from before backwards the most posterior being the last to appear, and til-- full nu i niter is reached when the embryo arrives at Sedgwick's Stage (I. The following is taken from the list constructed by S.-dgui''k t'r«>m all sources, including his own observations. As tinbibliography given by him is complete and easily accessible it is not repeated here, and the reader is referred to Sedgwick's monograph for reference to the original authorities.

/'. capensis: 17 pairs of claw-bearing ambulatory legs (Table Mountain. S. Africa ). /'. l>nl t',,11,-1 : is pairs of legs, of which the last pair is rudimentary (Table Mountain. S. Africa i. Sei|g\vi,-k ha- ••\ainined more than 1000 specimens from the ('ape. and haonly seen one specimen with more than 1<S pairs of leg-. This individual had ^0 pairs, the last pair being rudimentary. It closely resembled I', tin f /'mi ri , but differed in the number of lc'4-> and in certain other details (q. v.); Sedgwick regarded this form provisionally as a \ariety of P. Imlfmiri.

I', mosleyi: '2\ and :>:> pairs of legs : near AVilliamstown, S. Africa. The specimens with '2'2 legs w«-re two in number and were both females. They differed in certain other particulars from the form with i^l le->. but on the whole Sedgwick regards them as a Variety ..I' t he- >ame species. ()theispecies In-m S. Africa which have been less fully studied are stat.-d to ha\e lit, ^1 ;m,l -2-2 pairs ..f legs respectively." In all South African forms, irrespective of the number of legs, 'he generative opening is subterminal and is placed behind the la-t pair of fully developed legs (between the l.sth or rudimentary pair iii /'. l>nt/nin i). SI-.I.CWICK. j»p. 440 and 4.~>1.

Ill both o| these speeles tile geliel-ative opening IS between tile In all the Neotropical Species which have been at all fully examined, the number of legs varies among individuals of the same species. P. edwardsii: number of pairs of legs variable, the smallest number being 29 pairs, and the greatest number being 34. Males with 29 and 30 pairs of legs. The females are larger, and have a greater number of legs than the males. The new-born young differ in the same way. From 4 females each having 29 legs, seven embryos were taken which were practically fully developed. Of these, 4 had 29 legs, 2 had 34, 1 had 32. An embryo with 29 ami one with 30 were found in the same mother. An embryo, quite immature, but possessing the full number of legs, was found with a larger number of legs than one which occupied the part of the uterus next to the external opening. (Caracas.)

Peripatus demeraranus : 7 adult specimens had 30 pairs of legs; 6 had 31 pairs; 1 had 27 pairs. Out of 13 embryos examined, 7 have 30 pairs and 6 had 31. (Demerara.) Peripatus trinidadensis : 28 to 31 pairs of ambulatory legs. (Trinidad.) Specimens of other less fully known species are recorded as having respectively, 19, 28, 30, 32, 36 pairs of legs, &c. In the Neotropical Species, irrespective of the number of legs, the generative opening is placed between the legs of the penultimate pair. (SEDGWICK, p. 487.)

In connexion with the case of Peripatus, the following evidence may be given, though very imperfect and incomplete. 2. Myriapoda. CHILOGNATHA. Variation in the number of segments composing the body in this division of Myriapods cannot be observed with certainty ; for it is not possible to eliminate changes in number due to age, nevertheless the manner in which this increase occurs has a bearing on the subject. In Julus terrestris the number of segments is increased at each moult by growth of new segments between the lately formed antepenultimate segment and the permanent penultimate segment. At each of the earlier moults six new segments are here added: in Blaniulus the number thus added is four, and in Polydesmus? two fresh segments are formed at each of the earlier moults. In. each of these forms the number added is the same at each of the earlier moults. NEWPORT, G., Phil Trans., 1841, pp. 129 and 130.

CHILOPODA. The number of leg-bearing segments differs in the several genera of Chilopoda, but except in the Geophilidse the number proper to each genus is a constant character. For instance in Lithobius In Ceophiiidie, however, tintotal number of inoveable segments is much l.-u-^'.-r, r:ur_'in'_' fr«iin about .'55 to more than 200. Though not ehara.-t.-n-ti.- ..f -.-n.-iM, the iminlttT seems within limits to mark each particular It \\a- foil ml that male GeophiH have fewer segments

rather im-ater. In eiu'ht males the number varied between 96 and '.''.i; in eleven females, between 103 and 1<>7. <>f two female !ln some of the ( 'hiloj.oda' an increase in the number of segments afiei- i he lar\a hatches, hut the variations mentioned above are n-.-, u-.|eil ae o,-,-iiirinLr in fully formed specimens independently of s due to aije.] Iii tinton-^/n,^ cases, a tact which is often met in the Study «t Variation i- \\<M seen. It often happens that in particular LT--MITU or in particular -pecics. a considerable i-aiige of Meristic N'ariation i- found, while in do-ely allied forms there is little or none. Mxaiiijiles of t hiare seen in the variability of the (.Jenphilida- a- eoin|>areil with the other Chilopoda, and in the neo-tropical spedee of Peripcdus which vary in the number of legs, while P. 1ml t'nuri , for in-taiice. i> very constant. It will be noticed that in both th'- c.i — t he absolute numbers of parts repeated are con- -iderably higher in the variable than in the i-onstaiit forms. But though "ii'-h cases ha\e gi\-en rise to general statements that series -.f organs containing a small number of members are, as such, less variable than series containing more members, these statements require coii-iih-rable modification: for it is not difficult to give in-tanc,-- both in plant- and in animals, where series made up of a -mall number of members, shew great nieristic variability.

The bi •ai'injof t he-e ca-e- on the nature of M ei'ist if Repetition .md thconception of 1 lom, ,l,,^y will be cou-iJei-ed hereafter. II- re, ho\\e\er, it may be well to call attention to the fact that we have now before iicases in which various but characteristic num- b, i - ,,f legs or segments differentiate allied species or genera : that in a — liming the truth of the |)octrine of Descent, we have expressed our belief thai in each case the spi-cjes with diverse mim-

b,-i- are descended from some common ancestor. In the evolution of these form-, therefore, the number havaried : this on the one hand. On the other hand, in (n'njiliiltiti and in Peripabus, WQ see contemporary instances of the way in which such a change at its origin may be brought about. Though there are several things to be gained by study of these instances, one feature of them calls for attention now, namely, the definiteness of the variations recorded. The change from a form with one number to a form with another number here shews itself not as an infinitesimal addition or subtraction, but as a definite, discontinuous and integral change, producing it may be, as in Peripatus edwardsii, a variation amounting to several pairs of legs, properly formed, at one step of Descent. This will not be seen always to be the case, but it is none the less to be noted that it is so here.

Among Insects I know no case of such individual variation in the fundamental number of segments composing the body. Among Crustacea two somewhat remarkable examples must be mentioned, though it will be seen that both of them belong to categories very different from that with which we are now concerned. But inasmuch as they relate to the general subject of Meristic Variation they should not be omitted. 3. Carcinus maenas. The abdomen of these crabs consists normally of seven segments, including the last or telson. In the female the divisions between all these seven are very distinct. The abdomen of the normal male is much narrower than that of the female, and in it the divisions between the 3rd, 4th and 5th segments are obliterated. Males, however, which are inhabited by the Ehizocephalous parasite Sacculinn do not acquire these sexual characters, and in them there are distinct divisions between the 3rd, 4th and 5th segments. (Fig. 9 c.)

In male Carcinus nucnas inhabited by the Entoniscian parasite, Portunion, a similar deformity may occur, but is often very much less in extent, sometimes being only apparent in a slight alteration in the contour of the sixth abdominal somite. In specimens of Portunus, Platyonychits, Pilumnus and Xantho inhabited by Entoniscians, no change was observed. GIARD and BONNIER comment on the remarkable fact that the change in the sexual characters effected by Sncculina is greater than that resulting from the presence of Entonisciaus ; for since the latter are more internal parasites, preventing the growth of and actually replacing generative organs entirely or in part, it might have been expected that the consequences of their presence would be more profound. GIARD, A., and BONNIER, J., Contrib. a IV-tude des Bopyriens, Travaux de I'inst. zool. de Lille et du laboratoire zool. de Wimereux, 1887, torn. V. p. 184.

4. Branchipua and Artemia. A- it has been alleged that variation may be produe. d HI tl. tatiMn i.f the abdomen of these animals by changes in the a whi.-h th.-y live, it * -ary here to give the facts on which thistateitli this subject, that though not strictly cognate, unit »f tin- r\i.l.-nrr on this point also must be given. .1,7. H.I in. In- • es in the Mliiuty of the water in which the animals ted that Schniaiik. v. it-.-h observed the conversion of Hnuii-hijnix

int., I ,.f Artemia ...;//;ci into A. milhauxenii following upon the pro- i ..f tinwaters of a salt lake. Strictly speaking howeve* this is Srhmaiik. witsch. His story is briefly this: That tli.- salt Kiiyalm. 1 liy a diim into an upper and a lower part : the waters in th. latter In in/ satmated with .-alt, while the waters of the upper pan were less the 'lam ai i .1 the density of the lower waters to 8° Beuum. (=.i: ; u.'.li, au.l iii tinwater great numbers of A. xnUiin then appeared,

numbly hi\,!i • i...ii wa-h.-.l in from the upper part of tinlake, or from the hi. mil in After tinth.- .lain was made good, and the wat. is of the r lakiby evaporation became more and more concentrated, bring in the summer 1 , , ind lati i in that year tinsalt bewail to crystalli/.r out. i, until at tinend of the summer of 1*71 a large part of i had no caudal I'm-, thu> pir-i iitiii^.' tincharai '•• > ot A. milhausenii Fischer and Milin- Kdw. Tlir -iii-ci --ivr >ta!_'.-> of tindiniiimtion of the tail-tins and of the numhiTs of tinbristli - an- -hrwn in thr ligurr-i, with which all are now familiar.

, i."lur. .1 rx|.i-iiiin ntally by gradual concentration of water, inmr foi in i.-.-mbling .1. inilltiiiiai-nii. It was found also that if tinanimals without caudal tinwnr krpt in watrr which was gradually diluted, after M.iiie w..k- ;L j.air ..f conical ].romiiu-ncos, each bearing a single bri-tle, ap- It is fuitlier Mated that tinbranchial platesof the animals living in the more highly con.-, ntiat. d wat. r wenmaterially larger than those of animals living in

Schmaiil.i wit-ch K - on t.. -;iv that by artilicially brrrding Art, 'mid xnlinn in more and more dilut. d salt water he obtain..! a form having the characters of >• IIVMI me Input, and that li. i- this form as a new species of Ili-inii-lufiii*. lie ex].laintin- -lad ment thu-: In the normal Artrinin, the last \vhilr tbiingjiait in Hriini-lii fnm isdividid int. . two segnif-nts. He states that in his i. [union the condition of the la-t .-egmnit of the po-t-abdomen consti-

•ial dilTt-reiic.- betwe.-n Arti-mni and I'-rinichijnut, and that such by him from Art,-niin b;. i\r .lilution ,,f the watrr. A s.-cond distinction ily, while Itr.nii-liijiii* is not known to be so reproduced. As to the i .-litmn of hinew loim in tin- i. ~\- ot, S.-hinankewitscli had no evidence. iii-t. in.-, tb. • for tintype) and bv tingreater lei. the post- -ui. n. In the type the bi;-tles on each caudal lin air :'. m-rallv S — 12, and in

1 i ' thl point a ; ...... 1 deal of intrr. Ming evidence i- given iii Schmankeut as it does nut hear immediately on the ijiu-Mioii of the specific dii' •: b. . n introduced h< i. . var. a, 8 — 15, rarely more. Amongst specimens of var. a, as also among those of the type, specimens may be found having three, two, or even only one bristle on the caudal tin. The second antennas of the male are less wide in var. a .than in the type, and the knobs on the inner border are rather larger than in the type.

The variety b was found in pools of a concentration of 4° Beaume. It differs from the type in having the post-abdomen shorter in proportion, though the whole length is about the same. The number of bristles on the caudal fins is greater in the variety. The second antennas of the male are narrower in the variety than in the type, and bear a tooth and a thickening of the skin internal to the rough knoblike projections. But the most important difference characterizing var. b is the appearance of transverse segmentation in the last (8th) post-abdominal segment. This, according to Schmankewitsch, does not amount to an actual segmentation, but is really a transverse annulation, which may be more or less conspicuous, and suggests an appearance of segmentation. Schmankewitsch looks on this second variety as a transitional form between Artemia and Branchipus.

Before going farther it may be remarked that Schmankewitsch gives no figures of these varieties, except in so far as they are represented in the well-known series of sketches of the caudal forks with varying numbers of bristles. No analysis of the waters is given. It will be seen that two principal and distinct statements are made : (1) That A. milhausenii may be reared from A. salina by gradually raising the concentration of the water. ('2) That by diluting the water a division is produced in the last (8th) segment of A. salina: that this is a character, or, as Schmankewitsch says, the chief character, of the genus Branching.

First as to the relation of A. salina to A. milhausenii. The species milhausenii •was made by G. FISCHER DE WALDHEIMJ on spirit specimens sent to him, and the absence of caudal fins and bristles was taken as the diagnostic character. Fischer's figures are very poor, and indeed are scarcely recognizable: they are also incorrect in several points, giving for instance 12 pairs of swimming feet instead of 11. The description is also very imperfect. In the course of this he speaks of the male, saying that its second antennas are larger than those of the female, in which he declares the second antenna? may be sometimes absent. From Fischer's account it is quite clear that his material was badly preserved, and indeed, as Schmankewitsch says, specimens of these animals preserved with spirit only are of little use.

In 1837 EiTHKE2 gave a better figure of A. milhausenii ? from the original locality of Fischer's specimens. The tail, ending in two plain lobes, is shown. The male is not mentioned. The following analysis of the water is given : Other authors mention A. milhausenii, but there is, so far as I am aware, no special account of the male, or any material addition to the above. I will now give an abstract of such further evidence on this subject as I have been able to collect.

In the course of a journey in Western Central Asia and Western Siberia I collected samples of Branchiopods from a great variety of localities. Of these two consist of Branchipus ferox (Milne Edwards), one of Branchipus spinosu* (Milne Edwards), three of a species of Branchipus not clearly corresponding with any species of which a description is known to rne, and the remainder of Artemia. All the species of Branchipus collected are quite clearly defined both in the male and the female, and have certainly nothing to do with the Artemia. Of the latter some preliminary account may now be given, as the facts bear on Schmankewitsch's problem. Omitting those which were badly preserved and those which do not contain adults, there remain twenty-eight samples, satisfactorily preserved with corrosive sublimate, from as many localities. Of these, eight contain males, all of them having the

distinct! -Hii'i. It is difficult to speak with confidence as to the ; from tli.- f.-nial.- alone, but by careful comparison I can find -tru.-tunuhich ditT.-r.-ntiat.-s any of the remainder from the females ,,f many kind-, some being large salt lakes, while others were small pondi -T even pools. Th, -p. • -iiic gravities of these waters varied from 1-030 to tl,, i.t. 'Hi- ipecific gravities were measured in the field with u hvdr.,in« t. r nadii:/ to "i i:,, and on comparing these readings with the de-

t, and I think therefore tbat these rough readings are fairly tir\ ..f -odium cari.onate. so much that the water was strongly alkaline and the hand-. Thi> can generally be recognized on the spot in various .vitsch's work is that of the caudal fins. : • ige between the large fins with some twenty n t.. tincondition with no distinct lin or bristles. The following table tinniinihi r of bristles on the caudal fins, and this number is a fair guide to the size of the fins, large fins for the most part having many bristles and small fins having few. In the third column the range of this number in several individuals is shewn, and for this purpose only adult females bearing erjrjs in the ovisac are reckoned, as with sex and age there are changes in respect of the number of bristles.

Sulphuric anhydride S03 ... Carbonic anhydride CO., ... Lime CaO Oxygen equivalent to the Chlorine... Total solids in 1000 grams These analyses were undertaken for me by Mr H. ROBINSON, of the Cambridge University Chemical Laboratory, and my best thanks are due to him for the care with which he has conducted them. The table shews the great variability in the development of the tails and bristles. In specimens from the same locality there is generally great difference, and even the numbers on the two fins of the same individual are rarely the same. It will be seen that on the whole the forms with few bristles came from waters of high specific gravity, thus generally agreeing with Schmankewitsch's statement. This relation to the salinity is not however very close, but Schmankewitsch never asserted that it was. He frequently refers to the existence of individuals with tails in several conditions of degeneration in the same water, and especially (Z, f. w. Z., 1877, p. 482) he expressly states that in the original locality of A. milhausenii he found this form and with it several others intermediate between it and A. saliita.

It will also be seen in the Table, that the three samples, IV, XL and III stand out as having far more bristles than other samples from waters of equal specific gravity. Each of these localities was exceptional, and all belong to one class. Ill and IV were pools in the dry bed of a stream in the Kara Kurn, near the Irghiz river. They were close together, and must be joined in each spring. XL. was a pool in a somewhat similar dried stream-bed, coming down to the lake Tulu Bai in the district of Pavlodar. The conditions in these pools must be very different from those of the large, shallow, permanent salt lakes from which the other samples mostly came, and it is only fair to Schmankewitsch's case to remember that the water in such pools must be almost fresh during the early part of each summer.