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

plantar surfaces facing each other, and the tarsus R T is placed so that its dorsal surface is very nearly opposed to the dorsal surface of the normal tarsus LT, and the three tarsi thus flex almost in the same vertical plane. It is to be observed, however, that the tarsus LT is not actually in the same plane as the other two, but is a little deflected from it so as to flex rather more towards the posterior surface of the line than it would do if it stood actually as L'T' stands. This may be made clear by reference to the Scheme (p. 481) : for while the two extra tarsi are placed relatively to each other as if they were in the position D, the position of RT to LT is that which it would have if it stood in DDP.

In this species the middle tibia in the male bears one large spur, namely, the posterior one, while the anterior spur is rudimentary. PS in the figure represents FlQ_ 162 Aphodiug COHt(iminatuls, NO. 757, the large posterior spur of left middle tibia bearing extra parts. LT, the normal tarsus LT, while normal left tarsus. RT, LT, right and left i i ri /?P'*s'' extra tarsi. PS, normal posterior spur. 1U' ^ LP'S, Standing posteriorly of RT and LT corresponding with the single

and between the two extra sPur p* of normal, x, x, line of suture tarsi represents their two from the posterioi. 1surfac(, (Specimen the when it is examined from the anterior side, for upon that surface it is marked by a longitudinal ridge-like suture. This specimen fig. 13, and I am indebted to Dr Kraut/ fur an opportunity «>f "58. Galerita africana (Carab.) : (Fig. 163) right middle leg normal as far as the last tarsal joint, which bears three additional claws arising dorsally to the nmmul puii. The extra claws are three in number, two of them being small and standing at tinanterior border of the limb, while at the posterior border there is one claw of lai L-C r size. This larger claw is really a double structure, which is clearly shewn l.y the

en -i-nce of tu-n channels on its concave surface. Position of origin is therefore D, whilitiniui-liiiiitii.il i.l' tin- c.\tr:i pairs of claws to each other is about that required for the position DDA; fur the planes of the two pairs are not parallel but incline to each ntlii-r at an unit-' an^le. Specimen originally described by MOCQVKRYS, Col. tinoriii. , p. (i 1. I'm. *75!>. Pyrodes speciosus (Longic. ); having two supernumerary legs aiticulating \vith the thorax by a common coxal joint, which is distinct from the nixa of the left middle leg, but is enclosed in tinsame socket with it. In this remarkable case the normal le^ i.s complete, though slightly pushed towards the middle line. The socket in the mesothorax is enlarged posteriorly and dorsal ly so as to form an elongated, elliptical articulation, which lies obli'|uely, so that its ventral end is anterior to its dorsal end. The anterior and ventral end is occupied by the coxa of the normal leg, while the coxal joint of the two extra legs fills the space d'.r-al and external to it. Both are capable of being moved independently in the relaxed insect. The extra legs articulate with their coxa by a common double trochanter which has two apis's, from \\hich point the legs are distinct. Their position is dorsal and posterior to the normal leg, being practically that marked DP in the Scheme, and the relative positions of the extra le-s are very nearly those indicated for the Position DP. The leg nearest the normal leg is of course ji right leg in structure, and its plantar and a little of its structurally anterior surfaces anturned posteriorly. On the other hand, the remoter leg is a true left leg and the ventral surface of its femur is placed almost exactly horizontally. All three legs are complete, but they an- a little shorter and more slender than the middle leg

*700. Carabus irregularis $ ; left middle leg and right hind leg bearing supernumerary tarsal portions. In the left middle leg, Fig. 164, I, the 2nd tarsal joint is short and thick ; the 3rd joint is partially double, as shewn in the figure. One of its apices bears a tarsus of reduced size, and the other apex, which is postero-dorsal , bears a double tarsus having common 4th and 5th joints. The 5th joint of the latter bears two pairs of claws which curve ventrally and partly towards each other. The figure I shews the appearance from the ventral or concave side of the claws, while the figure II is drawn from the convex or dorsal side. The disposition and small number of the spines on the ventral side of the extra 5th joint shew that the ventral surfaces are partly suppressed, and in fact that the surfaces which are adjacent in the extra tarsi are in part ventral surfaces. This view is also borne out by the direction and curvature of the claws. Relatively to each other and to the normal the extra parts have nearly the Position DP.

FIG. 164. Carabus irrt'iiiilnria. I. Semi-diagrammatic view of left middle leg from antero-ventral surface. I1, the claws of normal left tarsus. ?-, /-, claws of extra tarsi compounded together. lhl, hair marking the dorsal surface between the claws. A, anterior. P, posterior. II. Dorsal view of apex of extra tarsus ;•//, Hi-, two hairs marking dorsal surfaces. III. Dorsal view of right hind leg. c'c', c"c", claw-like spines, perhaps representing extra claws. (Specimen the property of Dr Kraatz.)

The right hind tarsiis has the form shewn in Fig. 164, III. The 3rd, 4th and 5th joints are not fully separated from each other. Both the 4th and 5th joints bear extra parts, but their nature is obscure. The 5th joint is partly double, and the anterior part bears two shapeles* claw-like spines (c'c'). The 4th joint bears a similar pair of claw-like structures «,{' smaller si/.- (c"c"). Probably these should be considered as rudiment- ..f extra tarsi : but if tlii.s view is correct, it appears that tu.i extra tarsi ;uv present, arMn^ from different joints. For the loan i.f this -pe.-imen I am indebted to Dr G. KuAATZ, who first cleseribed and (inured it in l>-nt. . ,,t. Xt., 1S77, XXI. pp. -r>7 and 63, tijj;. 27.

7.; | Chrysomela graminia i I'll ytophagi) : the femur of the right middle leg bears a .-up. -111111111 mi -y pair of legattached to the posterior and dorsal side of its apex. At this iiuiiit tin-re is an articulation with which the single proximal part of the .-\ti.-t pair of til.i.-i- articulates. This piece, which is common to the two supernuiiieiaiy tibi;e, i- a -uli globular, amorphous mass from which the two tihi.-e di\.-rge. l.arh of the two tihi.-r- bears a complete tarsus, except that the most jK.-ti-rii.r has only one claw. In colour the two supernumerary tarsi differ from the noMiial, ln-iii;.' liinwn instead of metallic green, but the tibiae are normal in colour. I rom ti of the articulations and the arrangement of the pubescence, it is

rl.-ar that the surfaces <>t' the l,.gs which are naturally adjacent are constructed as jiu.-itt'rior surfaces, and the foinis of the two are complementary to each other, the himhiiost of the extia It ge li.-ing formed as a left leg, while the foremost is a right leg. A- thi -\ stainl, howe\t-r, the two tibia? are not in the same position relatively to tinhod\. t,ii tin foremost is placed normally, having its plantar surface turned downwards, but the hindmost is rotated so that its plantar surface is partially turned n>i-ii-<inlx The relative positions are nearly those marked DP in the Scheme, hut tht most posterior tarsus is more rotated than it should be according to that diagram. This condition may be to some degree connected with the presence of the ainoiphoiis giowth at the base of the extra tibia;. This specimen was kindly lent for description by 1'r Mason.

~(]'2. Pimelia interstitialis (Tenebrion.) : left posterior femur bears two supernuiiieiary tibiaarising from the postero-dorsal surface of its apex. These two are a pair, for the tibia nt ait -t to the normal tibia is a riyht tibia, the remoter being a left. The adjacent surfaces are chiefly anterior surfaces in structure, but the \entral Miifac. - are inclined to each other at an obtuse angle. The position of the extra legs is almost that marked DP in the Scheme, but the inclination of the Neutral sui faces of the extra legs is rather more acute than it would be in the I'ti-itioii I'l'. The. tarsi anall broken off. Specimen originally described by MOII..I i BT8, ''"/. iinnnii., p. 44, .////.

7li.'!. Acinopus lepelletieri (Carab.): two extra legs arising from posterior surface of base of femur of /. middle leg. From position it seems that the most anterior is the normal, but this i- doubtful. The arrangement is nearly that of Position DP, but as one of the femora is ci msti icted and bent, the relations are rather irregular. Specimen first de-cubed by MOCOUKKYS, Col. <niorm., p. 41,.%. "704. Silis ruficollis < Mularndrnii') : right anterior femur bearing a sujKTiminrrai v limb ( Fi.sj. 10.")). The coxa and trochanter normal. Tin' tfimir is of about twice the antero-posterior thickness of a normal frimir ami at its apex presents two articulations in the -ame horizontal plane. Of these the anterior bears a normal tibia ami tarsus, but the posterior bears an extra tibia which appears at Hrst sight to be a single structure. This tibia is more slender than the normal one and is provided with four tarsal joints, the terminal one being withered and without claws. Upon closer examination it is tiiimd that this extra tibia is in reality made of the /mxtrrinr surfaces of a /><n'r of tih'ne not separated from each other. In this case the morphological duplicity of the extra tibia is capable of proof. For, as shewn in Fig. IGo, II, the normal tibia is not bilaterally symmetrical about its middle line, the contrary the anterior surface is differentiated from the

posterior by several points. This may be seen in the spurs at the apex of the tibia, for the anterior spur (a) is long, but the FIG. 165. Silis ruficollis. No. 764. I. The right anterior leg seen from ventral surface. a, anterior, p, posterior. This figure was drawn with the microscope and is reversed. II. Detail of apex of tibia of the anterior or normal tibia, shewing a, the anterior, and p, the posterior spurs. III. Similar detail of apex of the tibia of the extra limb, shewing p' and p", two structurally posterior spurs.

posterior (p) is short (as is usual in the front leg of many beetles). The hairs on the surface of the tibia are also directed asymmetrically and the parting or division between them is not median, but is nearer to the anterior border (see figure). But in the extra part there is no such differentiation, and both surfaces are structurally posterior surfaces. The hairs part in the middle, and both spurs (p , ;/')are formed as posterior spurs. This extra structure is therefore made up of the two posterior borders of a right and a left tibia compounded together in Position P. (See diagram, Fig. 1G5, II.)

This specimen was found by Dr Sharp amongst a number of insects collected by myself in his company at Wicken Fen on Sunday1, July 26, 1891. Such a case taken in connexion with others (e.g. No. 801) makes it certain that many cases of supposed "single" extra appendages are really examples of double extra parts. 1 A day or two before, the manuscript of this part of the subject had been put by with the remark that no good opportunity of thoroughly investigating a case of " single" extra leg had occurred, but that it could scarcely lie doubted that traces of duplicity would be found in them. Considering the f-'ivat rarity of oxtra appendages in Insects, and remembering that even of the whole number very few are of the supposed "single" order, I have thought tho occurrence of this capture a coincidence of sufficient interest to be worthy of mention. Dr Sharp tells me that amongst all the beetles that have gone through his hands only one case of extra appendage (No. 755) was seen.

765. Scarites pyracmon (C'arab.). At base of posterior face of the t rochanter of lot't imriiial front leg, immediately above the cotyloid articulation was implanted an elongated lanciform joint. This joint was directed backwards and represented a pair of trochanters compounded by their anterior .surfaces. With each of the two apices of this double trocliantrr was articulated a complete leg, in all respects formed as an anterior leg. The two moved as a complementary pair. [Details given. This is one of the earliest and best described cases. . \-Mtvs1 in ijuotin.L,' it points out that the description and tigure plainly -In -w that tlie tv\o o\n-a legs were a pair, a right and a left, respectively. They uric in t'act a pair, arising from the posterior surfaces of the normal leg, and presenting their anterior surfaces to each other.] LKKKHVKK, A.. Gv&rin'a Mag. de Zool., 1831, Tab. 40.

76G. Geotrupes mutator (Lamellicorn) : two supernumerary limbs arising from ffiiiur of light anterior leg. Femur greatly widened, upon posterior border giving off a large prominence winch divides into two processes at right angles to each other. Karli of tin -. proi . --. •- hears u normal tibia and tarsus, but tbe foremost of these tibiie is shapid as a 1,'t't tibia, having its serrated border placed anteriorly, while the other extra tibia is formed as a rii/ltt tibia, having its serrated border placed tn>*t,-i-i«i-lt/. [The pair of limbs arise from the posterior surface of the normal limb and have their anterior surfaces adjacent, as in Position P.] FKIVALDSKY, J., Tt'rni. FUzetek., IMSI;, x. }>. 71 1. /'/.

7<ii. Pterostichus lucublandus ? (Carabidic): third tarsal joint of left middle leg at apex pic-, nts \\ide articular surface. On this stands a triple 4th joint, made up of a single anterior portion, bearing the rest of the normal tarsus and a posterior portion, double in structure, the two parts being completely united. The single anterior part of this Ith joint bears a normal 5th joint with claws. The double )io-terior part of the 4th joint bears a pair of separate 5th joints, each having a pair of rlav,s. Of these the anterior is perfect, but the peripheral part of the posterior •"tli joint is crumpled, so that its claws are twisted out of position, but at its base it -t.imlexacth a- the noimal .~,th joint, and as the '"'tli joint of the anterior extra tai.-ns, all three being in the same horizontal plane. These extra parts, therefore, anin the Position marked P in the Scheme and have the relations there indicated for that position. This specimen was kindly lent to me by Mr HARRINGTON, who.

'768. Ceroglossus valdiviae, Chili (Carabidse) : left anterior tibia bearing a pair of supernumerary legs. The tibia widens, and in its middle part gives oft' posteriorly and ventrally a wide branch having the form of a pair of tibial apices compounded together. The double tibia bears two tarsi (Fig. 100, .R', L') having a common proximal joint, but these have unfortunately been broken, two joints being missing from the one and three from the other. The legs, are ;i right and left as usual, and they stand in the relative positions m:irked \TP in the Scheme. This is a very simple and striking case, for the animal is of good size and the parts anwell formed. The two tibial spurs which are adjacent in the two extra tibiaancompounded so as to form a double spur \\ itli two points as sJiewn in the figure. As shewn for the Position

VPP in the Scheme, the compounded parts of the extra appendages, viz., the double tibia and the double first tarsal joint have two complete ventral surfaces inclined to each other at an obtuse angle, while there are only two halves of dorsal surfaces. FIG. 160. Ceroulossits valdivice, No. 768. Left anterior tibia with extra parts seen from the ventral surface. /., the normal left tarsus. R' and I/', the extra tarsi, compounded in their proximal joint. A, anterior. P, posterior.

Note that the anterior spur of the normal is curved and that the double spur representing the two anterior spurs of the extra tibial apices has thus a bifid point. (Specimen the property of Mr E. W. Janson.) Similarly there are two structurally posterior surfaces, but n<> structurally anterior surfaces, for these are adjacent and tindeveloped. This specimen was kindly lent by Mr E. \V. JAXSON. 769. Silpha nigrita (Heteromera): from right middle femur arises a pair of legs which are completely united as far as the apex of the la.M tarsal joint. The point of origin of the supernumerary limbs is on the anterior and ventral border of the femur. The form of the extra limbs is shewn in Fig. 167. The surfaces V and V are structurally vent ral surfaces. They are turned chiefly forwards, but are inclined to each other at an acute angle. The surfaces, therefore, which arc adjacent in this pair of legs, and which are consequently obliterated, arc chiefly the morphologically anterior surfaces and to some extent the dorsal surfaces. The plantar or ventral surfaces of the last tarsal joints arc inclined to each other rather more obtusely than those of the til>i;c. so that the curvatures of the two pairs of claws are very nearly turned forwards as well as away from each other. This is not fully brought out in the figure. The position of origin is about VP, but the claws are in Position VPP. Specimen tirst described and figured by Mon(>n:i;Ys, Col. anorm., p. 43, jiy.

Fi... 1(17. Siljili>i iii.iritu, No. 769. Eight middle femur bearing a compounded pair of extra legs. I', ventral surface of nearer extra leg. V, ventral surface of the remoter extra lef,'. (In Rouen Mus.) 770. Tenthredo SOlitaria (S,t\\lly): tibia of right middle leg divides in peripheral third to form two branches, of which the anterior bears i hr normal tarsus. Tlie posterior branch arises from the postero-ven tral surface of the normal and bears a double tarsus consisting of the posterior parts .,f a pair compounded in Position VP, almost exactly. Tibial .spin's as in Fig. IDC). Tlie compound tarsus has only 4 joints, the ">tli being apparently broken off. In Cambridge Univ. Mus., history unknown.

771. Telephorus rusticus ( Malacoderm ) : tibia of left middle leg dilated and somewhat deformed in its peripheral portion. It presents two apieal processes, the one anterior and the other posterior. The anterior of these bears a normal, backwanlly directed tarsus, but the posterior process bears t \\ o tarsi by separate articulations. The anterior of these two tarsi is directed t'or\\ards to t'ace the tarsus of the other apex, but the posterior tarsus is backwardly directed. [From its attitude it is dear that the middle of these tarsi is a structure complementary to one of the others, but there is no evidence to shew whether it is a pair to t lie anterior or to the posterior. Posit ion either VPP, or DAA, probably the former.] K ICAAT/, 1>* <it. ent, /?/., 1SSQ, p. 344, tig. 33.

Anthia sp. (Carabidffi) : left posterior tibia bearing two supernumerary tarsi. 'I'lif | "'.-ti-i "-ventral >id,. ,,f the apex of the tibia is dilated so as to form a triangular projrrti"ii. causing the jmint of articulation <>f the normal tarsus to be raised dp-wards. The projection bears two tai-M of which the posterior curves downwards and backwards, li.-iu^ fashioned as a left tarsus while the anterior curves forwards •iinl slightly upwards liein^ a ri'iht turtu*. These two tarsi ha\e unfortunately been broken but were j)resumabi\ coiiijilete. The whole apex of the tibia bears five spurs in-t' ad of two, but the relation of the >pms t" the separate tarsi was not clear. The

tarsi are very nearly in the Position YPP. Specimen very kindly lent by Mr E. W. Janson. 77*3. Julodis aequinoctialis (Buprestida?) : the extra legs arise from the posterior and ventral side of the base of the tibia of left middle leg. They are a pair, and are compounded together by their lateral and dorsal surfaces in such a way that the morphologically ventral surfaces of the two are almost in contact along the anterior border of the compound limb. The ventral surfaces here converge at an acute angle. The two extra legs are compounded together throughout the tibia- and first 4 tarsal joints. The 5th tarsal joints are free, but only one of them remains. The former presence of the other is only shewn by a socket. The normal tibia is constricted and bent at one point so that it does not stand in its normal position. The femoro-tibial articulation is rigid.

This is a case of a pair of legs compounded as in the position marked VVP in the Scheme but the point of origin is more nearly that of VPP. Specimen originally described by MOCQUKUYS, Col. anorm., p. 47 Jig. 774 Metrius contractus (Carab.) Esch. : specimen in which the middle left femur bears an incomplete pair of legs in addition to the normal one. The femur is of normal length. The tibia of the normal leg is articulated with the end of the femur as usual, but is somewhat shorter, stouter and more curved than the tibia of the corresponding leg of the other side. A supernumerary tibia arises from the posterior [and ventral ?] side of the femur a short distance within the apex, and is articulated with it by a separate cotyloid cavity; the two articular cavities for the two tibia; are confluent, being connected by a groove. The end of this tibia is dilated at its outer end, and bears two articular surfaces, one on each side ; with each of these surfaces, a complete tarsus is articulated, nearly normal in form but somewhat stouter than a normal tarsus. There are four terminal spurs to this tibia, two being below the outer tarsus and two being below the inner tarsus. [It therefore seems that this tibia is made up of parts corresponding with the ventral side of a right tibia and the ventral side of a left tibia, and it is hence probable that if the disposition of the claws of the tarsi had been examined, it would have been found that they too were a pair, one being a right foot and the other a left. Position probably VVP.] JAYNE, H. F., Trans. Amer. Ent. Soc., 1880, vm. p. 156, PI. iv. figs. 3

775 Aromia moschata ? (Longicorn) : right anterior coxa bearing a pair of supernumerary legs having trochauter and the proximal half of the femur in common. The normal leg and the extra ones were all somewhat reduced in size but \vure complete. The extra leg adjacent to the normal is a left leg. [From the figure it appears that the legs arose in the Position P, or VPP, and their relative positions seem to have been those indicated in the Scheme. Of course it is not possible to state this definitely without examination, but it is clear that there was at least no great departure from the position shewn in the Scheme.] It is remarked that in this specimen the right mandible was abnormally small. KOLBE, H. J., .\ntunr. Wochens., 1889, iv. p.

Two cases of double supernumerary tarsi require separate consideration. The arrangement in both of these cases departfrum that which is usually followed, but it will be seen that thenis considerable though imperfect agreement between the two exceptions. Both of these occur in the anterior legs of males of the genus Culatlms, and it happens that in the normal form the apex of the tibia presents a considerable modification from the simple structure of other beetles. This nioditieati<>n affects the anterior legs only, and is found in several genera «>f ( 'arabidae, being especially pronounced in Ciiltitlnt*.

In order to appreciate the nature of these cases it is necessary that the anatomy of the parts should be understood. Tinapex of the tibia in the simple form, e.g. the second or third leg in ('unilms, bears t \vo large articulated spurs. The two spurs are ventral to the articulation of the tarsus, and one of them is placed at tinanterior border of the tibial apex while tit. oilier is posterior. In these unmodified legs both spurs are placed at the same level in the limb, so that the bases of both are in the same transverse section (cp. Fig. 166). In the forms presenting the sexual modification, the anterior spur is of somewhat small size but occupies the same position relatively to the other parts that, it does in a simple leg.

The posterior spur however, which is large, does not stand at the same level on the tibial apex, but has, as it were, travelled up the tibia so that it stands at a considerable distance central to the apex, and instead of marking the posterior border of the limb it is placed nearly in the middle of the actual ventral surface. A long channel runs from the posterior spur to the anterior one, and the appearances suggest that the modified form is reached by a deformation of the original apical surface, which is twisted so that the posterior spur is thus drawn up into the secondary position. In the fore leg of a male ('arabus the beginning of such a change can be seen, hut in Pterostichus and especially in ('ninth us it reaches a maximum. The change may be briefly described by saying that a section to include the two spurs must be taken in a plane which is oblique to the long axis of the limb instead of transverse to it.

As a result of this modification the morphological surfaces "f the anterior tarsus of Calathus &c. have a peculiar disposition relatively to the same surfaces of the tibia when compared with other forms. Commonly the ventral surface of the tarsus is parallel to a line taken through the bases of the spurs, but owing to the rotation of the posterior spur into its secondary position this plane is here oblique to the ventral surface of the tarsus. These points will be at once evident if the front leg of a male Calnthiift is examined.