Foot evolution

ESPECIAL REFERENCE TO THE JOINTS

By HERBERT ELFTMAN AND JOHN MANTER

Columbia University

THE

current divergence of views concerningthe evolution of the human foot

isdue largely to a lack of accurate knowledge concerning thefunctioning of the

ape foot, and a concentration of attention on morphological features less intimately concerned withthe action ofthefootthan arc thejoints. Bymeans ofan apparatus for studying the instantaneous distribution of pressure in the soleofthe foot, we have been able to make an accuratecomparisonbetweenthe method of function of the chimpanzee foot and that of man (Elftman and Manter, 1935). In thel

present

paper we shall consider especially the joints

ofthe chimpanzee foot in comparison with Man, with regard totheir bearing

on the evolution of the human foot.

TIlE TRANSVERSE TARSAL JOINT AND THE LONGITUDINAL ARCHI

The transverse tarsal joint (mid-tarsal, Chopart's) lies between the

cal-caneus and talus on one side and the cuboid and navicular on the other

(figs. 1, 3). Movement in this joint allows the forepart of the foot, which is quite rigidly attachedtothenavicularand cuboid,tomovewithrespect tothe calcaneus and talus aboutan axis whichis shown in fig. 2. These movements we shall refer to as plantar-flexion and dorsi-flcxion about the transverse

tarsal joint. Since the axis of the joint in the normal position of the human foot is inclined at a considerable angle tothe horizontal, plantar-flexion with

the calcaneus held imnmovable results in abduction of theforepart of the foot. Dorsi-flexion results in abduction.

Whenmovementtakes

place

inthetransversetarsaljointwith the footon the ground, the tuber calcis andthe anterior

part

of the foot mustremain in contact with the substratum. Under these circumstances dorsi-flexion about

the transversetarsal joint results, notin abduction of the forepart of the foot, but in a loweringofthe joint and a flattening of the foot.

In thechimpanzeethetransversetarsal joint is freely movable, while in the human foot with a well-developed longitudinal arch the joint is relatively immobile. The human foot with its longitudinal arch (fig. 1B) displays the

samerelationships between theforepartof the foot andthecalcaneus and talus that we find in the chimpanzee foot when it is plantar-flexed about the trans-verse tarsal joint (fig. 1A). Weare justified in concluding that in the human foot the transverse tarsal joint has become fixed in a

plh.ntar-flexed

position.

Additional evidence for this concept is to be found in the prevalence of mobility inthe transverse tarsal joint in flat-footed humans. It seemslikelythat mobility inthe transverse tarsaljoint,rather than being a result of

flat-footed-ness, is more often a cause.

T T

JOINT,

,w

A

Fig. 1. Chimpanzee and human feet in medial view. A, chimpanzeefootin an arborealposition; plantar-flexion about the transverse tarsaljoint. B, human foot ontheground. C, chim-panzee foot on theground: pronation with dorsi-flexion aboutthe transverse tarsaljoint. T.T.transversetarsaljoint.

The evolution of the longitudinal arch in Man is consequently due to an

immobilization of the transverse tarsal joint in a plantar-flexed position. To provide a more complete explanation involves the determination of the

mechanism ofimmobilization. This cannot be done at present. It might be due to a shortening of such connective structures as the plantar aponeurosis, the

long plantar ligament and the plantar calcanco-navicular ligament. It could also be accounted for by changes in muscle action. It will be impossible to

decide this issue until we are able to settle by experiment the question as to

whether ligaments or muscles determine the limits of movement in the joints. The position of the axis of the transverse tarsal joint which wehave deter-mined differs from the position given by Fick(1911). There canbe no doubt

concerning the position of the axis in the chimpanzee. Since ourdeterminations

on mobile and juvenile human feet agree with the results in the chimpanzee,

wefeel that Fick must have been misled by movementsin other parts ofthe foot than the transverse tarsal joint.

A

B

C

T

2

U~~~~~

Fig. 2. Chimpanzee andhumanfeetindorsalview. A, chimpanzee footontheground. B,human footontheground. C, chimpanzeefootplantar-flexedabout the transverse tarsaljoint,asin fig. 1A. T. axis of transverse tarsaljoint. U. axis ofupper anklejoint. L. axis of lower ankle

joint.

Theimportance ofplantar-flexion orinversion about thetransversetarsal jointas oneof the factors in theevolution of thelongitudinalarch has already

been stressed by Keith (1923, 1929). The extension of his theory so as to assign primeimportance tothe tibialis anterioras an invertor of the foothas, however, been rendered doubtful by the studies of Fick (1911, 1931) and by our own determination of the position of the axis of the transverse tarsal joint. The tibialisanteriormaywell beapostural muscle, intimatelycorrelated with the evolution of theupright posture ofman, duetoits action about the upper ankle joint.

Attempts have beenmade toexplaintheevolution of thelongitudinal arch

plantar-flexion in the transverse tarsal joint, with the heel and anterior support of the foot in contact with the ground, does tend to bring the talar head more

directly above the calcaneus, especially since this plantar-flexion is usually

associated with inversion about the lower ankle joint. But the change in

position of the talar head is a

consequence,

not a cause, oflongitudinal arch production. The arch is a subtalar structure, depending on therelationship of the calcaneus to the anterior part of the foot.

Weidenreich (1921) recognises that plantar-flexion about the transverse tarsal joint has occurred in the evolution of the longitudinal arch, butconsiders this as being subsidiary and subsequent to the upward inclination of the cal-caneus. It seems obvious to us that the upwardinclination of thecalcaneusis anecessaryconcomitant of fixation of the transverse tarsal joint in the plantar-flexed position and cannot be considered as a cause of the plantar-flexion.

Certainly Weidenreich's argument that in Man the weight of the body is concentrated on the calcaneus and thus is responsible for its tilt is not sub-stantiated by our researches on the path of the resultant of pressure in the foot (Elftman and Manter, 1935). In the chimpanzee the resultant lingers for a longer time in the tarsal region than it does in Man, but the chimpanzee

gives no evidence of acalcaneal tilt.

Recognitionof the evolution of thelongitudinal archashavingbeen due to

plantar-flexion about the transverse tarsaljoint serves also toexplainthe fact that in the human foot the metatarsals are in line with the long axis ofthe

calcaneus, while in the ape foot, as seen resting on the ground (fig. 2A), the metatarsals diverge laterally from the longaxis of the calcaneus. If the chim-panzee foot is plantar-flexed about the transverse tarsal joint, with the heel and anterior support of the foot still in contact with the ground (fig.2C), it will

be foundthat the metatarsals in their new position are parallel to thelongaxis of the calcaneus.

THE LOWER ANKLE JOINT

In theankle there are two joints about which movement is possible. In the upperankle joint (talo-crural), the tibia can rotate backwards or forwards over the trochlea of the talus. It is the lower ankle joint (subtalar), however, to

which we shall first direct our attention. In this joint the articulation is

betweenthe'talus on one side and the calcaneus, cuboid and navicular on the other. This gives rise, according to the usual description, to the possibility of gliding movements in the joint. A moreilluminatingconcept is that ofDonitz, modified by Fick (1911), who demonstrated that, in spite of the bewildering

congeriesof articular surfaces involved, the movement in the loweranklejoint canbe quite accurately described as rotation of the entire subtalar portion of

thefootabout a line, the compromise axis of the lower ankle joint. Theposition of this axis is shown in figs. 2,

3,

4 and 5. When looking at the right foot from in front, a clockwise rotation of the foot about this axis would give rise to eversion, a counter-clockwise rotation to inversion.

60

Herbert

position by noting the relations of thetalus to the calcaneus, since the joint surface passes between these bones. The talus and calcaneus of the chimpanzee are represented in fig. 3A in the relative positions theyoccupy when on the

ground, but with the calcaneus tilted so as to make comparison with the other

figures easier. Fig. 3C shows these bones in the

positions

they occupy in the arboreal position offig. 1 A. It is obvious thatwhenon the ground the

chim-panzeefoot is everted about the lower ankle joint, while in the arborealposition illustrated a moderate degree of inversion is present. Ifthe human

relation-ships (fig. 3B) are now compared with the twofigures of the chimpanzee, it is

apparent that the usual human position in standing resembles that of the chimpanzee when moderately inverted.

A

B

C

Fig. 3. Anterior view ofthe talus and calcaneus,with theplane determined bythe axesof the transverse tarsal and lower anklejointsperpendicularto theplane of thefigure. A, chim-panzee,showingtherelationshipsof the bones in the terrestrialpositionoffig.IC,but with thewholefigurerotated outwardsoastobringthe calcaneus intoapositioncongruentwith thatoffig. 3C. B. human. C, chimpanzee,plantar-flexedlabout the transverse tarsal joint

withaccompanyinginversion about the lower anklejoint,asinfig.1A. U. axis of upper ankle

joint. T.,L.axesof transverse tarsal and lower anklejoint. A.L.P. antero-lateral process, its extentroughlyindicatedbythedottedline.

The normal human foot can neverachieve as

highly

everted a

position

as

that shown tobe the common terrestrial

position

of the

chimpanzee.

One of

the reasons for this is found

by

an

inspection

Of the anterior

portion

of the calcaneus of Man. Both in the anterior and lateral views of

fig.

5, and in

fig.

3B. therecanbeseenPa

bony

process, whichweshall callthe antero-lateral process of,the calc-aneus, which

prevents

extensive movements of eversion of the talus with respect to the calcaneus. No indication ofthisprocess can be

foundinthe calcaneus of either

chimpanzee

orgorilla. Itfurnishes oneofthe

most

easily

recognizable

diagnostic

cllaraclteristics of the human

type

of

calcaneuls.

RELATIONSHIP OF THE AX~ES OF TIDE TRANSVERSE-TARSAL

ANC

LOWER

ANKLE

JOINTS

The articulation between the talus and the navicular

presents

a

peculiar

problem, since movementinthis articulation maytakeplaceeither aboutthe axis of tigelower ankle

joint,

tcpabout e axis ofthe transverse tarsal joint, or

The Evolution

61

about both axes simultaneously. This would not be possible if the articular surfacewere notessentiallyspherical,with thetwo axesmentionedintersecting

atthecentreofcurvatureof thesphere. Since thetwoaxesdo intersectatthis centreof curvature, freemovement cantakeplacesimultaneously in the lower anklejoint and in the transverse tarsal joint.

The extent of the articular surface on the head ofthe talus is intimately correlated with the degree ofmovementpossible. The greater range of

ever-sion

possible

in the

chimpanzee

iscorrelated withtheextension of the articular

surface (fig. 4)furtherposteriorlyonthelateral side in thechimpanzee thanin

Man.

THlE UPPER' ANKLE JOINT

The talo-crural or upper ankle joint allows movement between the tibia andfibula, actingtogether, on oneside, and the talusontheother. We should consequently expectthat in the evolution ofthe human uprightposture, with

the knees held nearthe midplane of the body duringmovement, there would

be a remodelling of the tibia, fibula and talus. The differences between the chimpanzee and Man inthe tibia and fibula have been adequately treated by formerinvestigators.

A comparisonofthe talus of Man with that of the chimpanzee shows that the

troclhlea

in the human has been rotatedupon the bodyof the talus. This rotation isclearlyseenin thedorsalviews ofthebones(fig.4) by concentrating attentionupon theanglebetween theneck of the talus and the axis of theupper

anklejoint. This angle islargerin Man than in thechimpanzee. Thisdifference in angle has been noted by many previous workers. They, however, have interpreted itasindicatingashift inpositionoftheneck ofthetalus. That the neck hasnotchangedcanbeseenby notingitsrelationstothearticularsurfaces as seen in ventral view. Further confirmation ofour interpretation that it is the trochlea which has undergone rotation emerges from a comparison, in

posterior view, of the relation of the trochleatotheposterior calcancalarticular surface.

Themedial border of the trochlea ishigherinMan than in the chimpanzee,

ascanbe seenin theanterior and posterior views (fig. 4). This is alsocorrelated

with changesin the tibia.

TIHE METATARSO-PIIALANGE'AL JOINTS

That the five metatarso-phalangeal joints are differently oriented in

MIan

than in the chimpanzee and gorilla has been noted by many investigators,

including Morton (1922) and Weidenreich (1921). In fig. 6 the heads of the

metatarsalsareshown in projectionuponaverticalplane. It is apparent that

in the human there is not a well-developed transverse arch in this region, as

there isin the chimpanzee. Inaprevious paper(Elftman, 1934) the fact that thereis noconcentration ofpressure, suchas onewould expect atthe ends of a

functional arch, was taken to be an indication ofa lack of a transverse arch

thehumanfoot todistributeits pressure over this entireregion of the ball of the

foot when the heel is lifted, because of the disposition of the axes of the five

metatarso-phalangealjoints. In the chimpanzeethis isnot possible. The axes of the fifth and first metatarso-phalangeal joints are sharply inclined to the

ground. It is consequently impossible for the foot as a whole to bend in a

vertical direction inthisregion. When the heel of thechimpanzee and then the transversetarsal joint are lifted, pressure is transferred not to the region of the

metatarsal heads buttothe toes.

Theevolutionarychanges which have resulted in the human condition have

beendescribedastorsions ofthemetatarsals,the first in one

direction,

the four

lateral ones in the other. It is unquestioned that the distal portions of the metatarsals have undergone changes in which the bases have not participated.

THE JOINT AT THE BASE OF THE FIRST METATARSAL The opposability ofthe hallux in the apes contrasts greatly with its

per-manently adducted position in Man. Since this condition is apparent

exter-nally it has been seizedupon as the primedifference between the ape and the

human foot. There is no doubt but that in the evolution of the foot other

features aremoreimportant, butit is true that an adducted halluxprovidesa stronger anterior pillar for the longitudinal arch than would an abducted hallux.

The articular surface onthe first cuneiform of the chimpanzee is strongly

convex and faces medially. In Man the articular surface is more nearly flat and faces moreanteriorly. Some specimens of human cuneiform do, however,

possess a markedly convexsurface, theaxis of curvature being oriented as in

the chimpanzee, but the articular surface faces forward. There is no doubt whatever that thehumancondition could haveevolvedfrom thatfoundinthe

apes. Detailed evidence for this has beenput forward by Schultz (1930). THE BONES OF THE FOOT

The evolutionary changes in the joints have inevitably been accompanied

by changes in the bony elements. It isadvisable, therefore,tocomparebriefly

the moreessential features of the bones of the human and chimpanzee feet. The talus

Incomparing the talusof Man with that of the chimpanzeeitisnecessary

that thetwo bones beoriented in similarfashion. The usual method istouse

thetrochleaas abasis of orientation. This ishighlyunfortunate, sincewehave

seen that thetrochlea changes inthe course ofevolution. We shall therefore orient the two bones with the axes of the lower ankle joints parallel and the ventral articular surfaces in congruent positions.

When this is done, many ofthe differences which have been described as

existing between thetwo bones are seen tobe due tofaulty orientation. The generalsimilarityof the ventralarticular surfaces(fig. 4) isapparent. Itistrue

in Man. There is difficulty inestimating the importance of this, however, since the radius ofcurvature of these surfaces is shorter in small human feet than in large ones.

The rotation of the trochlea with respect to the rest of the bone and the increased height of themedial border of the trochlea in Man have already been discussed. The latter feature is of importance in studying the head of the talus. In the anterior view of fig. 4, it is noticeable that the greatest length of the talar head in the human makes a larger angle with the dorsal surface of the trochlea than it does in thechimpanzee. This has been interpreted as due to a torsion of the talar head. It is obvious, however, that here, again, it is the trochlea and not the head of the talus that has changed. The difference in

k~~~~~~~~~~

71 _ u, X.

DORSAL

VENTPRAL

ANTERIOR

POSTER/OR

Fig.4. Thetalus ofmanandchimpanzee oriented with theaxesof thelowerankle jointsparallel.

Upperrow: chimpanzee. Lowerrow:human. U. axis of upper anklejoint. L. axis of lower anklejoint.

extent ofarticular surface on the head of the talus when the chimpanzee is compared with Man contributes to this apparent torsion. But this we have already foundtobe dueto greaterpossibilitiesofeversion in the chimpanzee, not toa change in the axis of the joint.

The variations inthe anterior extent of the trochlearsurface in primitive tribes ofManhavebeen extensively discussed by anthropologists and need not be considered here.

The calcaneus

When the calcaneus is oriented by means of the axes of the lower ankle

jointand thetransversetarsaljoint(fig. 5), thereis astriking similarity between the chimpanzee and human bones. One of the chief differences to be noted is the appearance in Man of a new process, the antero-lateral process of the

calcaneus, which may be seen in anterior, lateral and dorsal views. The

significance of this in limiting eversion in the lower ankle joint has been

The similarity in contour of the facets for articulation with the cuboid is worthy of mention. Except for the additional area dueto the antero-latcral process, the human facet illustrated is characteristically like that of the chim-panzee. In some other human calcanei examined, the transverse groove which indents this facet is not so well developed. It is likely that variations in this groovearc correlated with the variability in the mobility of the transverse tarsal

joint.

The shape of thetubercalcanci, seen in posterior view (fig. 5), is extremely variablein Man. It is frequently quadrate in appearance, butvaries fromthis condition to one of ovate contour resembling more closely the chimpanzee. The calcaneus ofMan is relatively longer and stouter than that of the chim-panzee. In this character the gorilla resembles Man more closely.

I - (~~~~~~~~~~~~~~~~~D

-.ALP~~~~--ALPAL

"9'

Y

9\

XLj.4P'~t

X

V

<'tEV

~~~~~~~~~~~~~~A

\\\

DORSAL

L A

TERAL

AN

TER/IOR

FOSTERiOR

Fig.5. Thecalcaneus ofMan and chimpanzeeoriented with the axes of the lower ankle joint

parallel. Upper row:chimpanzee. Lower row: human. L. axisof loweranklejoint. T.axis oftransversetarsaljoint. A.L.P.antero-lateralprocessofcalcaneus.

Thesustentaculum talihas receivedtheattention of previousinvestigators,

some of whom have felt that its position in the human required extensive

explanation. A study of the anterior and posterior views of fig. 5 reveals no significant difference in the relationship of the sustentaculumto the rest of the

calcaneus between thechimpanzee andMan. Whatever differences it has with respecttothesubstratum as the foot is used must be due to the position of the entire calcaneus, not toany change inthe sustentaculum alone.

The trochlear process or peroneal tubercle is exceedingly variable in Man.

In the chimpanzee illustrated this process is more highly developed than it is

inthehuman, but in another specimen of chimpanzee calcaneus it approaches closer to the human condition.

Other tarsalelements

The navicular, cuboid and second and third cuneiforms of Man are longer than those of the chimpanzee, when compared with the combined length of

tarsus andmetatarsus. This lengthening is balanced to a certain extent by the shortening of the four lateral metatarsals. The euboid of Man has a larger surfaceforarticulationwith thecalcaneus, coveringin partthe anteriorface of

the antero-lateral process of the calcaneus. The differences in relative pro-portionsof the second and third cuneiforms cannot be discussed here; they may be correlated with a more important feature, the shape of the first cuneiform. The surface on the first cuneiform which articulates with the first metatarsal has already been discussed.

Fig. 6. The heads and thearticular surfaces of the bases ofthe metatarsals, projected upon a vertical plane. The articular surfacesof the bases are shown in dotted lines. A, human. B, chimpanzeeinnormal terrestrialposition. C,chimpanzeewithfoot plantar-flexedabout thetransversetarsal joint.

Metatarsalsand phalanges

In addition tothe torsion of the heads of the metatarsals with respect to

their bases, there is another difference of importance between Man and the chimpanzee. This is the shortness of the four lateral metatarsals and their phalangeal series in the human. Schultz (1930) has demonstrated that in the

course ofevolution the firstmetatarsal and its phalanges have undergone no

marked changeinlength, when measured with relationtothe restof thefoot,

but that there hasbeen an actual reduction of the four lateralmetatarsals and

theirphalanges, especiallythe second phalanges. It is significantthat in the

gorilla, which is more terrestrial in its habitat than is the chimpanzee, the lateral phalanges are shorter.

THE MECHANISM OF FOOT EVOLUTION

The contribution of the present paper lies chiefly in a clarification of the structural differences between theapefootand thehuman. Ananalysisofthe

mechanism of foot evolution must interpret these structural differences from

twopointsofview: first,the changesin the developmentalprocess which lead

todifferences inadult structure;and secondly, the differences in functionofthe adultstructureswhich determine whether they shall be judged fitorunfit by the environment in which they struggle for existence.

The first of these problems has been studied by numerous investigators, notably in recent years by Straus (1927). It would be well worth while to re-analyse thedescriptive embryological data available intermsofthe structural changes noted in the present paper,with a viewto finding,not evidence for recapitulation, but some clue to thedevelopmental factors involved. Wefully realise that an analysis into such factors as differential growth rates is but a

preliminarysteptowardadefinitiv-eanalysis in morefundamentalterms, such

as genie differences.

The second problem, that ofselection by the environment for procreation ofthosestructureswhichare functionally advantageous, isin some ways more

easily studied. We candetermine, inafashion,the changeswhich have taken place in the environment. In the case at hand, we at least know the

environ-ments in which the chimpanzee and Man live. We cantell, by studying the

structuresexperimentally,why they havesurvivalvaluein the formsnowliving. If we knew more definitely the environments which surrounded the stages

intervening between theape ancestor and Man,wecould make afair estimate

as to which types of structure would have had survival value during this critical period.

SUMMARY

The fundamentalsimilarity in architecture demonstrated inthe

comparison

ofthe foot of thechimpanzee with thatof Manleavesnodoubtas tothe evolu-tion of the human foot from that of an ape. But superimposed upon this fundamental similarity are important differences which allow us to makethe

following list of themostimportant changes which the foot has undergone in its

evolutionfrom ape to Man:

1. Stabilisation of the transverse tarsal joint in a position of plantar-flexion. Thisaccountsforthelongitudinal arch andpartially for the alignment ofthe lateral metatarsals with the long axis of the calcaneus.

2. Preventionofextensive eversion in the lowerankle joint, partlyby the

appearance ofa new process, the antero-lateralprocess of thecalcaneus.

3. Rotation of the trochlea ofthe talus with respect to the body of the bone, and raising of the medial border of the trochlea, correlated with the

changed position of the tibia in upright locomotion.

4. Torsion of the metatarsal heads, allowing flexion in the metatarso-phalangeal joints perpendicular to the ground. This eliminates the anterior

transverse arch as a functional factor and allows the distribution ofpressure over the ball of the foot.

5. Permanent adduction of the hallux, forming a strong anterior support

for the longitudinal arch.

6. The following changes in proportion: shortening of the four lateral metatarsals and their phalangeal series; increase in relative length of the

cuboid, navicular and first and second cuneiforms; and increase in stoutness and length of the calcaneus.

Although the gorilla resembles Man more closely than does the chimpanzee in the relative shortness of the lateral digits, it shows no indication of the more fundamental changes, such as those in the transverse tarsal joint, which are essential for the development of the human condition.

The fact that the human foot, adapted as it is for walking on the ground, bears a closer resemblance to the ape foot as used in arboreal than in terrestrial locomotion may be regarded as another evidence of Man's arboreal ancestry. Itwould also suggest that the essential features of Man's footwere acquired at an early stage of his terrestrial existence, rather than after longapprenticeship on the ground. It would be perfectly possible, however, for a human type of foot to evolve from either the chimpanzee or gorilla foot as now constituted, since the structures which, by modification, can give rise to the human structure, arestill intact.

REFERENCES

ELFTMAN, H. (1934). "A cinematic study of the distribution of pressure inthe human foot." Anat.Rec. vol.LIX,pp.481-91.

ELFTMAN, H. and MANTER, J. T. (1934). "The axisof the human foot." Science,vol.LXXX,p.484. (1935). "Chimpanzeeandhuman feetinbipedal walking." Amer.J.phys.Anthrop. vol. xx, pp.69-79.

FICK, R. (1911). Handbuch derAnatomie und Mechanik der Gelenke. Dritter Teil. Jena: Gustav Fischer.

(1931). "tberdieBewegungen und die MuskelarbeitandenSprunggelenkendesMenschen." S.B.preuss.Akad. Wis8. S.458-94.

KEITH, A. (1923). "Man'sposture." Brit. med. J.pp.669-72.

(1929). "Thehistory of the human foot and its bearing on orthopedic practice." J. Bone Jt

Surg.vol.XI,pp.10-32.

MORTON, D. J.(1922). "Evolution of the human foot. Pt. I." Amer. J.phys. Anthrop. vol. v, pp.305-36.

SCHULTZ, A.(1930). "The skeleton of the trunk and limbs of higher primates." Hum. Biol. vol. II,

pp.303-438.

STRAUS,W. L., Jr. (1927). "The growth of the human foot and its evolutionary significance." Contr. Embryol.Carneg.Instn,vol.XIX,pub. 380,pp.93-134.

WEIDENREICH,F. (1921). "DerMenschenfusz." Z.Morph.Anthr.Bd.XXII,S.51-282.