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

74-0. Melolontha vulgaris ? (Lamellicorn) : right posterior femur bears a suj)fi-- mimrniry pair of limbs having a double tibia in common. The supernumerary partannitln-r smaller than the normal ones. [The position of origin and symmetry, according to tintigure, must have been approximately V.] KOI.KE, H. 74], Carabus perforatus - : from thr ventral or plantar surface of the 5th tarsal joint of Irt't hind leg project an extra pair of claw-joints compounded in Position V. ca.-li braiing a pair ..f daws, set back to back. This is a diagrammatic case, well and clearly d.-sciili.-d by aSHUBS, .!/«'/<>•//•. (',,/,,.;<., 1835, p. 54, Tab. IX.

742. Feronia i Pterostichus ) miihlfeldii $ (Carabidse) : left middle tibia bearing two supernumerary tarsi arising by a common proximal joint (Fig. 15(3). As in other cases of supernumerary tarsi arising from the tibia, the apex of the tibia itself is really a triple structure, containing parts of the apices of a pair of tibia: in addition to the normal apex. This is shewn by the presence of three pairs of spurs, &c. The additional parts are in this case anterior and ventral to the normal apex and a complementary pair. All three ancompletely blended together, forming in appearance a single apex. The relations of the three component parts are almost exactly those indicated in the Scheme for the Position VAA.

FIG. 156. Pterostichus miihlfeldii, No. 742. Semidiagrammatic iv]>rrs.-m;itifii of the left middle tibia bearing the extra tarsi upon the antero-ventral border of tli.- apex. L, the normal tarsus ; E, the extra right ; I/ the extra left tarsus, property of Dr Kraatz.) The two extra tarsi (R, L'} arise by a common proximal joint of double structure having two complete ventral surfaces inclined to each other as in the Position VAA. Peripherally to this the two tarsi are separate. The tarsus which is nearer to the normal tarsus is perfect, and stands in the schematic position, second joint of the remoter arises in the position shewn for \ AA. but its apex is slightly shrivelled and in consequence tinremainder of this tarsus, though perfect in size and form is thrown a little out of position. This specimen was kindly lent to me by Dr KRAATZ, and was originally described and figured 1»\ him in Dent, ent. Zt., 1877, xxi. p. 56, fig. 21.

Aromia moschata $ (Greece) (Longicorn) : righl tibia enlarged at apex bearing anteriorly a supernumerary pair "t tar-i. The widened apex l)cars three supernumerary spurs of which tinmiddle i. DC is thicker than a normal posterior -pur. 'I'lii- i~ no doubt a double spur representing tin- t \vo posterior -purof tin- i-xtra tibia-. The <>\ her two extra spurs are ordinary anterior -pur-. Tinrelative positions of these spurs are exactly those marked \'AA in the Scheme. Of the extra tarsi 3 joints only remain and the two tar.-al series are so closely compounded that superlieialK they seem to form one tarsus only. In their tir.-t joints the inclination of the ventral surfaces to each other i- at an <n-,it,- angle. thus departing from the Scheme, but in the -econd and third joint-, where they are more separate from h nt her, the inclination is at approximate!} the same angle as that of the lines joining their respective spurs. Specimen in (ieneial Collection of the British Museum.

74 1- Carabus graecus ^ : troehanter of right middle leg bears a supernumerary pair of legs having trochanter, femur, tibia and 1st tarsal j"int ciiiiiiiiuii. The coxa of the normal leg is enlarged and the trochanter has two heads, of which the anterior belong to the extra pair of legs. 'I'he t'eiuur of the extra pair is a single piece but is morphologically double, presenting two *trn<-tii milt/ anterior surfaces and two structurally ventral surfaces, the latter being inclined to each other at an angle of aliout 1 •_'(). From the apex of this t'eiuur there arises a double tibia, also < ...... posed of two anterior and two ventral surfaces. This fact is

especially dear in the case of the tiliia and is proved by the arrangement of the spines and spurs. In a normal tibia there are two spurs, "ne posterior and one anterior, and the posterior spur is longer than the anterior. Now in tin's tibia there are three spurs, two shorter ones at either margin of the apex, and one hniiji-r one with a bifid point between them, which is dearly therefore n pair of posterior spurs not completely separated from each other. This view of the structure of the double tibia is equally evident from the arrangement of the remainder of the spines on its surfaces. In it the inclination of the ventral surfaces i- aliout the same as in the femur, but is perhaps rather more acute. The 1 >t tarsal joint is similarly a double structure. Its apex presents two articulations, but while the posterior bears a complete •! jointed continuation, the anterior bears only a single aborted joint, from which possibly some portion has been detached, but this is not certain.

The relations of the parts are a little obscured by the fact that the normal tibia is slightly bent. The double part of the trochanter lies \ery nearly anterior to the single part but it is also somewhat dorsal to it This ^ives tn the base of the double femur a trend dorsalwards: but from the base the femur Curves vent ralwards so that the nett result is that its ape\ is actually \entral to the apex of the single femur when both limbs are extended. This curve of course gives the femur an abnormal form which is increased by the fact that it is perceptibly shorter than the single femur. No\v the relative position of the pair of extra limbs is that marked VAA, and as it stands when extended the apex of the double femur and the peripheral parts of the double limb stand in the Position VAA with regard to the single limb ;

but as has been mentioned, by the curvature of the double femur its base is somewhat dorsal to the single limb. This specimen was very kindly lent to me by Dr L. VON HKYDKN and was first described and figured by him in Deut. ent. Zt., 1881, xxv. p. 110, fig. i'">. Eurycephalus maxillosus (Lougic. ): right anterior femur divides at base into two parts, of which the posterior bears a normal leg. The other part of the femur is bilaterally symmetrical, being made up of the anterior surfaces of two femora, for both sides present the same convexity (Fig. l.">7), neither being flattened as the posterior surface of a normal \\\->\ femur is. With the apex of this joint articulates a bilaterally symmetrical tibia of extra width, bearing a 1st and 2nd tarsal joint, each of nearly double width.

The 2nd tarsal joint bears two 3rd tarsal joints, which are both much wider than the normal 3rd joint of the tarsus. (This is exaggerated in the diagram.) One of these in 1891 bore a perfect terminal joint with a pair of claws ; but the terminal joint and claws of the other side were gone, though Mocquerys' figure shews that they were originally present. Mocquerys' statement that "la cuisse ante'rieure du cote droit se bifurque des son origine en deux branches ayant cJtacnne le volume d'une cuisse normal e" is misleading, as suggesting that the two femora are similar, while upon closer examination they are seen to be dissimilar. Here a pair of extra legs arising from the anterior surface of the normal limb, are compounded together as in the position marked A in the Scheme. Specimen originally described by MOCQTEHYS, Col anorm., 1880, p. 54, ji<j.

Eros minutus (Malacoderm) : righl anterior tibia divided at apex, forming two apices (Fig. 158). The posterior apex bears a normal tarsus. The anterior apex bears a double tarsus having the first three joints simplr (3rd being enlarged The 4th joint is of nearly double width and bears peripherally two claw-joints each with a of tin--.- it was clear that they are a pair. When extended the ill!.-'- plantar surfaces are ii"t truly iu a horizontal plane, as they

Aleochara mrrsta tStaph.i: middle left tibia has two articulations at apex. p.i-t. nor bears a tar>us normal in form but without claws. The anterior bears an extra tar-al s. riewith a ]>nir of rudimentary terminal joints, each having a pair .if claw-. ( »f this double tarsus the :ird and 4th joints are not distinctly separated. Tinparts are in I'o.-ition A. Specimen kindly lent by Dr MASON. 74S. Meloe proscarabacus J i Hi teromera). The apex of the femur of right hind leg i- < xt. iidi '1 "ii the aiiterioi side so as to form a second apex in the same horixontal plane. With this second apex articulates the common head of a pair of extra tibiaeach bearing a complete tar>u-. As ir-ual they are a right and a left respectively. Ih. two extra legs are twisted out of their natural position so that they turn their ventral surfaceupwaid-. The tibia which in origin is remoter from the normal tibia is moreo\ei bent over the nearer tibia MI that it stands actually nearer to the noi mill tibia. In this way the morphological relations are obscured, but nevertheless on tracing the. ventral surfaces up to th. point of articulation with the femur it is • •leai that they arise in the normal position and that they have the relations marked in the Scheme for the 1'o-itioii A. \\hicli is their position of origin. As this case is a -omewhat oli-cure one. I may add that I »r Sharp, who has kindly examined this -p. cimen. gives me l<a\e to state that he concurs in the above description. This is the -pecimeii de-cnhed by VON HKVHKN, />•/.<, isi-iii, ix. ]). 7lil and by MOCOI-KKYS, <'<il. .iiKTni. p. ">-2, ii;i.. and was kindly lent to me by Dr L. VON HKYDHN in whose possession it remain-.

( Hi. Cetonia opaca i l.amell.) : [right fore leg bears a pair of extra terminal tarsal joint- v.ry nearly in Position A, arising from 4th tarsal joint. All the claws are tinned \entralwaid-. but those of the extra joints are turned away from each other a- well a- downward-J. Mm QU1 RYS, I.e., p. til. /i.'/. Prionus coriarius $ ( Longirorn ), having parts of a supernumerary pair of tarsi arising from the middle right tarsus, and also a •similar double Mrueture arising from the posterior

right tarsus (Fig. 159). This is a very important case as a clear FIG. 159. Prionus coriarius, No. 750. I. Apex of tibia of right hind leg with its tarsus. II. Similar parts of right middle leg. (The property of Dr vmi Hrv.len.) PS, AS, posterior and anterior tibial spurs belonging to the normal IC.L'. l'S't A"S", the spurs of the extra legs, all structurally anterior spurs. A, anterior surface. V, ventral surface. (The property of Dr von Heyden.)

illustration of the mode in which double supernumerary limbs may be compounded together so as to closely simulate a single limb. Each of the extra parts in this case in the original account was described as a single extra limb, but as will be shewn, each is really composed of parts of a complementary pair. ( 'uses of this kind suggest very strongly that other cases of supposed single extra limbs are really instances of double extra limbs in which the duplicity is disguised.

Jiiml //'<f i Fig. 1")!', I), the tibia is dilated towards the :i|i'-\ which presents dor-allv two emarginations instead of one as usual. On tinventral aspect of the apex there are two whole spurs /'>'. J "N" and a double one AS, A'ti', between them. Tin--.- -purs give tinkey to the nature of the structure. The proximal laical joint L;'i\'es oft' a process on its anterior side and i- then riiiitinued to bear a normal termination as shewn in the figure. The prore-.-, IVoiu the first tarsal bears a second tarsal t'ron i whirh the termination has been broken off. The extra parts are ;is in the figure, being covered vent rally from edge to edge with papilla-, and having n<> loiujit "tliniil cleft in the middle line like t he normal tarsus.

Looking at these tarsal joints alone, the real nature of the extra parts does not appear, for the anterior and posterior surfaces of the normal tarsi are not differentiated from each other, and hence it is not possible to say of what parts the supernumerary limb is made up. Fortunately, however, the tibia! spurs are normally distinguishable from each other, for the anterior spur is a short spur while the posterior is a long thin spur. Now the spurs present in this case are firstly one long posterior spur /JN, and then three short « uti'rinr spurs, of which two are united for part of their length AX, A'S'. The extra spars are thus both iinti'i-ior sfHirs, that of the extra tarsus which is nearer to the normal being united to the normal anterior spur. Hence this case is a case of a supernumerary pair of appendages compounded together in the Schematic Position A, having the posterior surfaceadjacent and suppressed.

Rnjlit in'nl, II,' lt'ij. (Fig. ].")!), II.) Iii this case there would have been more difficulty in making out the real nature of the parts ; for in the normal middle leg the anterior spur is not so much differentiated from the posterior one as it is in the hind leg: but having this case for comparison it is easy to see that this also is a rase of a pair of appendages similarly compounded in Position A. This case differs from that of the hind leg in the fact that the parts are not so fully formed, and especially the anterior spur of the nearer extra tarsus is scarcely separated from the anterior spur of the normal. By turning the specimen over in the li-hi however, its form can be made out to be that shewn in the figure. When the specimen was received by me the parts present were as shewn in the figure, but when originally described by vox HEYDEX there was a third joint in the extra appendage which was small and elongated, and to all appearance it was the original termination and nothing had been broken off. For the loan of this specimen I am indebted to Dr L. vox HEYDKN, \\ ho originally described and figured it in Deut. ent. Zt., 1881, XXV. p. 1 10, ////.v. -11 and 2N.

In the two following cases there was nothing to differentiate the normal limb from the two supernumeraries, and the Position may either have been P or A. 751. Foenius tarsatorius (Ichneumon): tibia of left posterior leg bears a pair of supernumerary limbs. This is rather a remarkable case by reason of the great similarity in the modes of origin of the three limbs, whence it is difficult to determine positively which is the normal one. The tibia divides into three parts which lie in a horizontal plane and are separate from each other for about ^ of the length of the tibia. Of these the anterior is a good deal more slender than the other two which are similar and about of normal size. The middle of the three is shewn by its spurs to be a ritiltt limb. Each bears a complete tarsus The ventral surface of the most anterior tibia is horizontal while those of the other two are not quite so, but converge downwards ut a very obtuse an^'le. From this fact, and from the equality in si/e between them, it seems probable that the two posterior limbs are the supernumerary pair. The Position is therefore very nearly P or perhaps A. This speciniun was described by Mr HARRINGTON in Can. Knt., IS'.lO, p. 1'24, who was so kind as to lend it to me.

752. Agestrata dehaanii (Lamellicorn) : the coxa of the right anterior leg has two articulations, one anterior and the other posterior. With the anterior there articulates a single trochanter, bearing a normal right leg. The posterior articulation bears a large structure which is composed of two trochanters united together. This double trochanter bears two legs and is placed in such a way that the two do not lie in the same horizontal plane ; but the posterior extra leg is in the same horizontal plane as the normal leg while the anterior extra leg is wedged out towards the ventral surface, between the normal leg and the posterior extra leg. The posterior extra leg is a normally shaped right leg having its structurally anterior surface forwards as normally. The anterior extra leg is fashioned as a left leg and the surface of it which is structually anterior faces backwards towards the other extra leg. These two are therefore a complementary pair, having their structurally anterior surfaces adjacent: all three legs are normal and similar in form, size and colour. [Specimen kindly lent by Mr E. W. Jauson.]

*753. Cerambyx scopolii (Longicorn.) : pair of extra legs arising from the coxa of the right anterior leg. As this is a remarkably simple and perfect case it will be well to describe it in some detail, as it will serve to illustrate the arrangement of such cases in general. A normal leg of such a beetle as Cerambyx consists of coxa, trochanter, femur, tibia and four tarsal joints. To a proper understanding of the mode of occurrence of the extra legs in this case it is essential that the forms of these parts and their mode of movement with regard to the body and to each other should be accurately known.

Of the large, irregularly pear-shaped coxa only the hemispherical factis seen from the exterior. It is chiefly enclosed by embracing outgrowths from the sternum, forming a socket in which it can lie rotated like a ball. Upon its broad, exposed surface it is itself hollowed out to form a socket for the ball of the t >•(><•// a nf IT. For our purposes it is necessary to find some means of distinguishing the anterior face of the coxa from the posterior face. The structure which at once enable* u- to do this is the process (Fig. 160, /»), which goes down from the coxa to embrace the neck of the ball of trochanter and lock it into its socket. Now in the case of an anterior leg, this process is posterior to the trochanter (but in a middle or hind leg it is anterior to the trochanter). The next point to be considered is the position of khe femttr. The

lemur it-elf is Haitem-d anteni ]ni-teriurly. having t\vo broad surfaces, morphologically anterior anil ]»>-terior, and two narrow surfaces which an- e\ten>i>r and tlf.xnr Mirfaee-, or morphologically dorsal and ventral. I'.y rotation of the mxa tinwhole leg may assume a great variety • if portion-, and it i- thuof tlie utmost consequence that the nature of the surface- In- truly recu^ni/ed. If the front leg be placed with the ,.. process ..f ooza l..,-kii^ j,, ti,L. tr,,clianter. *rp, alp, corresponding processes !<•!• the extra trochanters. (From a specimen belonging to M. H. Gadeau de Kerville.)

femur at right angles to the body it may either be placed so that the ventral surface is downwards, or by rotation of the coxa through 90' the broad posterior surface may be downwards. The rotations of the middle and hind legs are complementary to this. In the abnormal specimen the extra pair of legs arise from the anterior side of the normal coxa, forming with it a solid mass and preventing its free rotation in its socket, so that the normal leg can scarcely be moved from the first position with the ventral surface downwards. The common coxal piece is about half as large again as the normal. Posteriorly it bears the trochanter of the normal leg, which is of full size and of proper proportions. The process of the coxa locking in the ball of the trochanter is posterior, as in the normal front leg.

Anteriorly the legs SL and SR articulate with the coxa by separate trochanters. Each is separately closed in by a process of the coxa, sip and srp, respectively. Of these processes that of the leg SR is posterior, but that of SL is anterior. Hence the two legs are complementary to each other, and SR is a ric/ht leg while SL is a left. This complementary relation is maintained in all the other parts of these legs. In size the two extra legs are rather more slender than the normal leg.

It was explained in the introduction to the subject of supernumerary legs that the relations of form between them depend upon the surface of the normal leg from which they arise. Here the point of origin is chiefly anterior to the normal leg, but is also slightly nearer to the extensor or dorsal surface of the coxa. This is not at first sight evident owing to the rotation of the normal leg due to the great outgrowth from its anterior surface; but nevertheless if the plane of the ventral surface of the normal femur were produced, it would pass ventrally to the ventral surface of the remoter extra leg SR, and therefore this leg is morphologically dorsal to the leg R. The positions of the extra legs are approximately those of the Scheme for the radius marked DA A, and while the surfaces of SR are parallel to those of R when both are extended, the surfaces of SL are inclined slightly to them as in position DAA. In the enlarged Figure III the coxa is rolled forwards so as to exhibit the relation of images between R and SL, and the figure II shews the coxa rolled back to shew the similar relation between SK and SL.

For the loan of this beautiful specimen I am great!}' obliged to M. Henri Gadeau de Kerville. 754, Harpalus rubripes (Carabichr) : left posterior tibia bears a supernumerary p:iii of tarsi. The apex of tbe tibia is widened and presents two articulations, of which the posterior bears a normal tarsus. The anterior articulation bears a pair of complete tarsi having proximal joints compounded. The two extra tarsi are a complementary pair, the posterior being fashioned as a ri/ilit. The surfaces adjacent in these two tarsi are structurally posterior surfaces, but they are a little supinated, so that the ventral surfaces are also partly turned towards each other. The position of origin and the relations of the surfaces to each other are almost exactly those which are

7 ."">. Chrysomela banksii . l'h\ tophagii : right hind tibia bearing an extra pair of tar-i. The ImnliT of ili'- tibia which cm i .--] Kinds in position to the ventral or flexor i. r «if tinnoimal tibia i> co\eied with the hairs which characterize it iu the iioimal limli; but tb.- opposite bonier of this abnormal tibia is similarly covered with bail-. sh. -\\ing that tinanterior parts of at least two tibiaarc included in it. A iigid pioc, •-- piojictiiom the wide apex of the tibia. Upon the inner side of tinpi. ic. sa i- the artinilatii.ii t'..r the tai-u-, which from its direction and position appear- to be the iicninal tat>u- . >f the limb. Outside the process articulates a -lightly Mnall. -r tar-iis, wliicli from its form and from the plane in which it moves /• it tar-u-. flexing away from the normal one. At a point slightly external to tin- i- the third tiir>u-. which is again a ri;iht tarsus and moves in a plane complementary t.i the middle one. The two are therefore a pair. The position of origin i- anteii.,1 and .im-al. being neatly that marked DA, but the relative positions of the extra tat -i or< approximately I'l'A. As to the nature of the tibial process I can make n ; me. 1 1- ig. ir, 1 . i

KM. Kil. Clii-i/xnnii'lii liinikfii. No. 7 •">•">. View of right hind tibia from posterior -111 tact'. A normal right hind tibia i- shewn for comparison. (From Proc. Zool. Sot-.. -]'• oimeo tinproperty of L>r D. Sharp.) Thi- -i'eeini, ti is the property of I>r Sharp, who was good enough to lend it to me. It was briefly described and figured by me /'. '/. . .s.. Is'.in, p. 583, but 1 was not at that time aware of the c»mplementaiy relation existing in these cases and failed to notice the >ome\\hat incoii-picuous differences which are evidence of it in this case.

7">'i. Hylotrupes bajulus (Longic.): right middle tibia bears a supernumerary pair of limbs having jiroximal jiaits in common. From the antero-dorsal surface of the ba-e of tinnormal tibia, there arises a slender tibial piece which is not so long •innormal tibia and bears no spurs. At the apex of this supernumerary tibia. which i> doubtless a double -tincture, articulate a pair of tarsi having their tii -t and ,-econd joint* compounded together. After the second joint the two tar>i -epaiate from . adi other an 1 each bears a pair of claws. The relative position of the two tai-M when tli.-\ separate from rarh other is almost exactly that marked DA. li -hould lie meiiti.inid that the -upei nmnerary parts central to the 3rd tarsal joints annot fully foime,i. u ing deiicieiit in thickness, and the transverse separation between the 1-t ami 'Jiid t:n>al joints is incomplete. Specimen first described by MO,V,IM:\-. Col. anonn., l-^sn. p. :,:!. Hg. i am indebted to Dr L. vox HEYDEN for an oppoitunity of examining it.

'7.37. Aphodius contaminatus^ (Lamellicorn.) : left middle tibia li.'.'irin^ two "U|MTiiii]iiri-ai-y tarsi \vhidi stand very nearly in the l>"siti"ii DDP, being rather m-aivr to D. The relative positions are shewn in Fi.^- l(i-- Tlvarti.-ular surface at the apex of the til>ia is extended aL-n-' an el, moated pnieess which projects on the d..r-al side .if the tdiia. Tj ..... this extension «>f the ape\ articulate two extra tarsi. They -tand with their ventral or