The Evolution of the Mm. Interossei in the Primate Hand 0. J. LEWIS Department o f Anatomy, St. Bartholomew’s Hospital Medical College, Charterhouse Square, London E.C.1 ABSTRACT Various conflicting views relating to the phylogenetic history of the interossei are reviewed. The primitive mammalian (marsupial) precursors of these muscles are shown to present a bilaminar arrangement: a dorsal layer of four bipennate abductor muscles (inserting into a proximal phalanx) is overlaid ventrally by a sheet of ten flexores breves, grouped in pairs, and inserting as wing tendons into either side of the extensor aponeurosis of the corresponding digit. The homologues of these muscles are identified in the hands of representative Primates including Homo. The dorsal abductors become the dorsal interossei proper; the flexores breves become the palmar interossei, which are therefore frequently more numerous than the four found in man. Certain of the flexores breves show a tendency to merge with those subjacent abductors with which they insert. Thus, the descriptive human dorsal interossei are composite muscles resulting from the amalgamation of a flexor brevis with a dorsal interosseous proper. Comparative morphology is shown to provide a logical basis for the understanding of the extensor apparatus of the human fingers. The interosseous muscles have long presented morphologists with certain perplexing features. These palmar (or plantar) adductors and dorsal abductors are characteristically grouped about the middle digit, but the axis of the primate hand (or foot) varies among species and even intraspecifically. It may be shifted to the second or even the fourth digit. Any such shift apparently requires a drastic reshuffling of the muscle insertions, contrary to the usual principles of myological evolution. Moreover, the muscles vary numerically, some Primates having as many as six, or even seven, in the palmar series. The supernumerary ones, however, although part of the same palmar sheet as the remainder, are so disposed as to act as abductors, closely related to and reinforcing the normal dorsal abductors. Any satisfactory phylogenetic hypothesis must explain equally shifts in axis and variations in number. The various explanatory hypotheses proposed all derive essentially from Cunningham’s (1882) studies on the marsupial hand and foot, which established the primitive mammalian pattern of the intrinsic musculature (fig. 1A). Cunningham demonstrated the presence of a basic trilaminar arrangement of digital muscles consisting of a dorsal layer abducting ANAT. REC., 153: 275-288 from the middle digit, an intermediate layer of bicipital flexores breves with metacarpal origins and insertions either side of the corresponding digit, and a superficial layer of four adductors (Mm. contrahentes). He held that the human condition could be derived from such an arrangement by assuming that the dorsal abductors become the dorsal interossei, certain of the flexores breves become the palmar interossei, the remainder disappearing (fig. l B ) , and the first only of the contrahentes remains as the M. adductor pollicis.’ Cunningham’s view was adopted by Thane (1892). A second, widely held, hypothesis derives from the work of Ruge (1878) on the development of the pedal interossei. He claimed to have shown the development of both plantar and dorsal interossei from paired muscles lying initially on the plantar aspects of the metatarsals (in the position of Cunningham’s short flexors), those bellies destined to become dorsal interossei later reaching their definitive position by migration into the intermetatarsal spaces. Windle ( 1883) reported similar findings for the human ___.- 1 Cunningham had earlier (1878) supposed that the contrahentes layer provided the palmar interossei, a view he modified later. His critjcs (e.g., Campbell, ’39) have sometimes failed to conslder the later, more plausible version. 275 276 0. J. LEWIS Fig. 1 A diagrammatic representation of the various suggested theories on the evolution of the Mm. interossei. A, the arrangement of the dorsal and intermediate muscle layers in the primitive mammalian hand (the contrahentes layer is not shown); the four bipennate dorsal abductor muscles are labelled Abl, Ab2, Ab3 and Ab4; the ten flexores breves are stippled and indicated by Arabic numerals (the marginal members of this series, the first and the tenth, are shown in darker stippling); the metacarpals are indicated by Roman numerals. The shading conventions shown in A are retained in all other parts of the figure and a muscular component which is believed to disappear is represented by a broken line, as is the primitive site of a muscle which has undergone a supposed migration; the arrows indicate the direction of such postulated movement. B, the view of Cunningham (1882). C, the view of Windle (1883). D, the view of Forster (‘16). E, the view of McMurrich (‘03). 277 PRIMATE MM. INTEROSSEI hand (fig. 1C). This hypothesis recognized no derivatives of Cunningham's dorsal abductor layer in the human hand and was supported by Campbell ('39), Keith ('48) and Haines ('50). Its attraction seems to be its provision of a facile mechanism whereby to explain shifts in the palmar or plantar axis, the axis becoming dependent solely upon which individual muscles undertake the dorsal migration. A third hypothesis interprets the dorsal interossei as composite muscles, resulting from the amalgamation of a dorsal abductor with a short flexor - in each case one of those short flexors which Cunningham supposed to have disappeared (fig. 1D). This view received support from a detailed comparative study by Forster ('16) and was deemed helpful by Bunnell ('42) and by Eyler and Markee ('54) in elucidating the human condition. Earlier a somewhat similar view was proposed by McMurrich ( ' 0 3 ) , modified, however, by his rather unorthodox views upon the phylogeny of the thenar and hypothenar musculature. Most morphologists believe that the marginal members of the flexor brevis series (the first and tenth) have, as their human homologues, the M.flexor brevis pollicis and the deep portion of the M. opponens minimi digiti ( St. John Brooks, 1886); McMurrich, however, believed that the thenar and hypothenar muscle groups are entirely derived from the primitive tetrapod flexor brevis superficialis. Inherent in this conception is the difficulty of having to account for all ten of the short flexors in the remainder of the palm. McMurrich's solution, summarized in figure lE, postulated that the first and tenth flexores breves have become the first and fourth human palmar interossei, and that the first dorsal interosseous muscle, like the fourth, incorporates not one, but two flexores breves. - RL Fig. 2 The extensor apparatus of the human left ring finger (anularis). The terminology used is that of Landsmeer ('49, '55); alternative terms used by Haines ('51) are shown in brackets. DB, deep belly (superficial belly) of fourth dorsal interosseous; DTL, deep transverse ligament of palm; ET, extensor digitorum tendon; LI, lamina intertendinea; L3, tendon of third lumbrical; PI, third palmar interosseous; PP, palmar pad (with the attachment of a slip of the metacarpophalangeal collateral ligament to its margin); PT, medial tendon (phalangeal tendon) cut short of its insertion to the proximal phalanx; RL, oblique band of retinacular ligament (link ligament); SB, superficial belly (deep belly) of fourth dorsal interosseous; TL, transverse lamina (deep layer of hood ligament); WT, wing tendon (side slips). 278 0. J. LEWIS This involves the supposition that in each case one of these two short flexors has transferred its insertion across an interdigital cleft. Phylogenetic considerations apart, renewed interest has recently developed in the insertions of the Mm. interossei in man, and in the detailed anatomy and function of the digital extensor apparatus, in consequence of which a much more precise knowledge of the human arrangement has emerged. Figure 2 is taken from a dissection performed with the hindsight provided by the work of Bunnell (’42), Braithwaite, Channell, Moore and Whillis (’48), Landsmeer (’49, ’55) and Haines (’51). Certain minor details have been omitted from this diagram : thus, Haines (’51 ) described an additional thin superficial layer to the hood, and Landsmeer (’49) described the retinacular ligament as having an additional transverse band. It is clear that, usually, a dorsal interosseous muscle divides into two bellies, the more dorsal belly inserting on the proximal phalanx and the more ventral furnishing a tendon which crosses more superficially to a dorsal insertion into the extensor expansion. As Salsbury (’37) first pointed out, however, the relative contributions to these two insertions vary with the different dorsal interossei: the first inserts solely into the proximal phalanx; the other three insert additionally into the extensor expansion, the phalangeal insertion being greatest for the second, least for the third, and intermediate for the fourth. Landsmeer (’55) moreover showed that the tendon of the dorsal belly reaches its phalangeal insertion by passing deep to the transverse lamina, whilst the tendon of the ventral belly passes superficial thereto to attain the extensor expansion. He also showed that the first dorsal interosseous may also have a ventral belly inserting into the transverse lamina or even into the wing of the extensor expansion. Various authors (Forster, ’16; Salsbury, ’37; Landsmeer, ’55) have stressed that the palmar interossei (like the ventral bellies of the dorsal interossei) are inserted entirely into the extensor expansion, reaching it by passing superficial to the transverse lamina - yet nevertheless many textbooks still describe a partial phalangeal insertion as the normal arrangement. The interesting generalization which emerges from these considerations is that the extensor apparatus of each finger displays symmetrically disposed wing tendons, derived either from a palmar interosseous or from the ventral belly of a dorsal interosseous (and reinforced on the radial side by a lumbrical). The division of the dorsal interosseous muscle into dorsal and ventral bellies is interpreted (Salsbury, ’37; Haines, ’51; Landsmeer, ’55) as a functional modification related to the different insertions and actions of the two bellies - the phalangeal tendons acting on the metacarpophalangeal joints, the wing tendons acting on the inter-phalangeal joints, as described by Stack (’62, ’63). These descriptive findings lack that basis of understanding which comparative morphology often provides. Yet there are many features suggestive of a common basic primate plan. As already noted, it is universally accepted (fig. 1 ) that the human palmar interossei are derived from Cunningham’s intermediate palmar layer of short flexors. There can also be little doubt that the supernumerary so-called palmar interossei, found in many other Primates, are of similar derivation. Henceforth in this paper all these muscular units in subhuman Primates will be referred to as Mm.flexores breves (and the individual bellies will be identified by reference to the numbering in fig. 1 ) ; this nomenclature is preferable to “palmar interosseous,” since this latter term, borrowed from human anatomy, carries with it functional implications and notions of homology not applicable throughout the primate series. In this context it must be significant that the human dorsal interossei are double-bellied, each belly - dorsal or ventral - having a characteristic mode of insertion. It is herein suggested 2Note how Landsmeer (’55) and. Haines (’51) described these bellies as “superficial” and “deep”; confusingly, however, they used these terms in opposite senses, the one author considering arrangements from a dorsal viewpomt, the other from the palmar aspect. 3 They should perhaps more properly be called Mm. flexores breves profundi, the term applied to their homologues in the amphibian and reptlle hand. These lower tetrapods also present a flexor brevis medius and superficialis, represented i n the mammalian hand by the more superficial muscle and tendon layers. 279 PRIMATE MM. INTEROSSEI that in fact the human dorsal interossei are composite muscles derived from one of the primitive two-headed dorsal abductors with a phalangeal insertion, compounded with a flexor brevis, and it will be shown that a detailed analysis of the insertions in relationship to the digital extensor apparatus in man and other Primates is wholly confirmatory of this conclusion. OBSERVATIONS Trichosurus vulpecula This species shows a close approximation to the basic mammalian arrangement described by Cunningham and illustrated schematically in figure 1A. The four bipennate abductor muscles of the dorsal layer are orientated about the middle digit, and are overlaid ventrally by the sheet of flexores breves. The marginal members of this latter layer, being comMATERIAL AND METHODS ponents of the thenar and hypothenar The present hypothesis is obviously groups, are beyond the scope of this disstrengthened if it can be shown that in cussion; thus eight muscles (numbers two the primitive (marsupial) condition the to nine in fig. 1) remain for consideration. digital dorsal abductors have a proximal The first of these is related to the ulnar phalangeal insertion whilst the flexores side of the first metacarpal, the last to breves of the intermediate layer insert the radial side of the fifth metacarpal. into the extensor expansion. The classical The remainder are grouped in three pairs literature affords no information on this arising from the central three metacarpoint but implies that the muscles of both pals. It is preferable here to speak of layers insert into the phalanx or into a pairs of individual muscles rather than to metacarpo-phalangeal sesamoid. This ap- adopt the customary designation of single bicipital muscles, since the individual parent detail, which however has signifi- members of each pair are, in fact, encance in the present context, was re- tirely separate entities. examined by a dissection of the hands of The tendons of the dorsal abductors the common brush-tailed possum (Tri- merge with the surfaces of the metacarpochosurus vulpecula) and the Virginian phalangeal joint capsules and insert into opossum (Didelphys marsupialis). the bases of the proximal phalanges. The The interossei were carefully dissected tendons of the flexores breves pass to a in the hands of the following primate spe- more dorsal insertion and are totally atcies: chimpanzee ( P a n satyrus), white- tached to the margins of the correspondnosed monkey (Cercopithecus nictitans), ing extensor tendons (not the base of the weeper capuchin monkey (Cebus nigrivit- phalanx as hitherto described) here formtutus) and white-thighed colobus monkey ing wing tendons. The wings of the extensor apparatus are not, therefore, (Colobus polyhomos vellerosus). The Mm.interossei were dissected in 25 uniquely human characteristics but anhuman hands, particular care being taken cient mammalian features. (Description to determine the precise details of their of the contrahentes layer is irrelevant to the present discussion. ) insertions and the relationship of these to the digital extensor apparatus. The first Didelphys marsupialis palmar interosseous was exposed from the The general muscle pattern reflects that dorsal aspect by reflecting the radial head described for Trichosurus but with some of the first dorsal interosseous from its modification of the muscle insertions. The origin on the first metacarpal. In many flexores breves are but partially continued cases the insertions of the other interossei as wings to the extensor tendons, for, as were most easily examined by longitudi- they skirt the metacarpo-phalangeal sesanally splitting the extensor digitorum ten- moids, they gain an attachment thereto. don of the separated finger, and laying The fourth dorsal abductor muscle has a either half (with its attached transverse double tendon of insertion, slips passing lamina and wing tendon) away from the to both sides of the fourth interdigital metacarpo-phalangeal joint . cleft where they find insertion into the 280 0 . J. LEWIS The pattern of insertion may best be considered by first describing those situations where there are tendons of both a dorsal interosseous and a flexor brevis, e.g. the radial side of the medius (fig. 3 ) . As in man, a strong transverse lamina passes from each side of the extensor tendon to invest the metacarpo-phalangeal joint and attach to the palmar ligament. It constitutes a useful morphological landPan satyrus mark. The tendon of the dorsal interosAmong the Primates this species may seous commences on the volar surface of usefully be considered first, since the the muscle and the fully constituted tenmuscular components corresponding topo- don passes deep to the transverse lamina graphically to the human interossei are to insert into the base of the proximal here clearly defined and that blending of adjacent muscle bellies which has occurred repeatedly in other Primates (and even in members of other mammalian orders) is here minimal. There are four dorsal abductor muscles, similar in disposition to those in marsupials, each arising from a pair of adjoining metacarpals, and abducting digits two, three and four from an axis formed by the middle digit; these muscles are usually called dorsal interossei. On their palmar surface is a series of fusiform muscles, each arising from a single metacarpal, and clearly homologous with the flexores breves. Seven such are present, corresponding to numbers three to nine inclusive in the primitive palm (fig. 1A); these are the muscles commonly described as palmar interossei. Most authors (Hepburn, 1892; Sonntag, ' 2 3 ) have recognized only six of these muscles, failing to observe the most radial member of the series (number three) Forster ('16), however, recognized the full complement of seven in a specimen of Troglodytes niger. (In the present specimen muscle number three was an entity clearly separated from the underlying dorsal interosseous but perhaps, as will be shown for the homologous muscle in man, there is here individual variation.) Of these muscles, numbers four, seven and nine 3 The muscles and extensor apparatus correspond in position to the ulnar three onFig. the radial side of the middle finger (medius) human palmar interossei, and will be of the right hand of Pan satyrus. The radial head shown to be homologous therewith: mus- of the second dorsal interosseous has been cut cles three, five, six and eight are, how- from its origin on the second metacarpal; the ever, intimately related, both at origin and flexor tendons and the second lumbrical muscle are removed. DI, second dorsal interosseous; FB, insertion, to the subjacent dorsal inter- flexor brevis; L2, tendon of the second lumbrical; ossei. other lettering as in figure 2. contiguous sides of the proximal phalanges of the fourth and fifth digits. This is a common, if not canonical, insertion for this muscle in marsupials, and was noted earlier by St. John Brooks (1886) in Didelphys and Trichosurus (although not observed in the present specimen of Trichosurus vulpecula). PRIMATE M M . INTEROSSEI phalanx. The tendon of the flexor brevis passes superficial to the transverse lamina and is inserted entirely into the extensor aponeurosis, forming its wing tendon, and is joined here by the second lumbrical tendon. Although the ventral flexor brevis has the descriptively more dorsal insertion, its tendon does not cross over that of the dorsal interosseous, owing to the habitually flexed position of the metacarpo-phalangeal joint. This description is also applicable to the insertions on the ulnar side of the medius, the ulnar side of the anularis and the radial side of the index. In this last digit the separate identity of the flexor brevis is emphasized by the well-marked aponeurotic surface of the first dorsal interosseous; the flexor brevis is represented proximally by a slender tendon of origin only, but its muscular belly extends far distally onto the dorsal interosseous tendon, before giving rise to a rather feeble tendon which inserts in the usual fashion as a wing tendon. On those sides of the digits lacking a dorsal interosseous insertion - the ulnar side of the index and the radial sides of anularis and minimus - the arrangements are as shown in figure 3 , but with the omission of the dorsal interosseous : the flexor brevis (i.e. a palmar interosseous) passes superficial to the transverse lamina to insert exclusively as the wing tendon of the extensor apparatus. In the specimen dissected there was no muscle corresponding to the human first palmar interosseous - number two of the flexor brevis series. However in this situation St. John Brooks (1887) described a fibrous band in the chimpanzee and a muscular belly in the orang and gibbon. All the flexor brevis tendons, as they lie superficial to the transverse lamina, are invested by a thinner more superficial layer of the hood. As noted above, a similar thin superficial lamina of the hood ligament occurs in man. The flexores breves have a fusiform appearance in norma palmaris, often with a small proximal tendon of origin attached to the metacarpal base: the deep aspect of the fusiform belly, however, gains appreciable origin from that ventral part of the metacarpal shaft which is free 28 1 of dorsal interosseous attachment and an intervening fibrous septum emphasizes the separate identity of the two muscles. The insertion of the fourth dorsal interosseous shows a departure from the general plan elaborated above, recalling the common double insertion of the homologous marsupial muscle. In addition to the usual abducting phalangeal tendon passing deep to the transverse lamina on the ulnar side of the anularis, a second slender tendon arises from the aponeurotic palmar surface of the muscle, and disappears beneath the transverse lamina on the radial side of the minimus. This latter tendon does not, however, reach the base of the proximal phalanx here, but gains final attachment to the metacarpal head. Cercopithecus nictitans In essentials the arrangement of the intrinsic musculature is comparable to that described for Pan, seven, discrete, flexores breves, and four dorsal interossei being present. Again the former insert by wing tendons into the extensor aponeurosis and the latter by abducting tendons into the bases of certain of the proximal phalanges. (The fourth dorsal interosseous possesses only a single abducting tendon, the supernumerary chimpanzee tendon to the radial side of the minimus not being elaborated here.) The transverse lamina has not achieved the high degree of differentiation observable in Pan, but the more extended position of the metacarpo-phalangeal joints results in the dorsal interosseous tendons being crossed superficially by those of the flexores breves, as these proceed to their more dorsal insertions. Cebus nigrivittatus The hand shows the usual four dorsal abductors (dorsal interossei). There are, however, eight members of the flexor brevis series, seven occupying the same locations as those in Pan and Cercopithecus, together with an additional one on the ulnar side of the first metacarpal. This last muscle belly, homologous with the human first palmar interosseous, OCcurs (apparently variably) in other monkeys and has been noted by St. John 282 0. J. LEWIS Brooks (1887) in Cercopithecus, Macacus, Pithecia and Hapale. Those four members of the flexor brevis series which are homologous with the human palmar interossei (corresponding to numbers two, four, seven and nine in fig. 1 ) all exhibit a typical insertion as wings of the extensor aponeurosis. In this species the transverse lamina is variably differentiated regionally, but where well-marked it lies as usual deep to the palmar interosseous tendon. The bellies of the remaining flexores breves (numbers three, five, six and eight) are somewhat merged with the subjacent dorsal interossei. The tendons of insertion of the flexores breves and the dorsal interossei on the radial sides of index and medius show a mutual crossing similar to that noted in Cercopithecus, as they pass to their respective insertions into the wings of the extensor aponeurosis and the proximal phalangeal bases. The insertions on the ulnar sides of medius and anularis are, however, modified but the mechanism of the change here is unmistakably clear and elucidates certain parallel human specializations. In each instance the partially united bellies of the flexor brevis and the dorsal interosseous give rise to their own tendons which cross as usual alongside the metacarpo-phalangeal joints : both tendons, however, insert into the wing of the aponeurosis, there being no phalangeal insertion. The tendon of the dorsal interosseous has, in effect, relinquished its phalangeal attachment and merged with the deep aspect of the overlying flexor brevis tendon, which thus carries its insertion distally into the aponeurosis wing. Such a process is, of course, a common feature of myological evolution elsewhere. Colobus polyhonzos vellerosus The usual four dorsal interossei are present plus seven flexores breves, comparable to the seven described for Cercopithecus. The three flexores breves corresponding to the three ulnar human palmar interossei (there is none corresponding to the first of man) insert as usual by wing tendons into the extensor aponeurosis. The transverse lamina is poorly developed. The bellies of the re- maining four flexores breves are even more intimately merged with the related dorsal interossei than in the preceding species. Despite this, in each instance the tendons of the two morphological components retain their separate identity and cross each other as in other monkeys. The basal simplicity of insertion noted in Cercopithecus is not, however, here completely retained: on the radial side of the index the flexor brevis wraps over the tendon of the dorsal interosseous (as it does in Pan and sometimes in Homo) but does not attain the wing, the whole mass inserting on the phalangeal base, and on the ulnar side of the medius some of the dorsal component is annexed by the flexor brevis tendon, a very slender phalangeal tendon being alone retained. On the radial side of the medius and the ulnar side of the anularis the insertions are uncomplicated, with tendons crossing, as usual, as they pass to their respective insertions into the proximal phalangeal bases and the extensor aponeurosis wings. Homo The human interossei fit easily into the general pattern already described for other Primates. Any apparent uniqueness investing them is largely the result of a traditional terminology which ascribes to man no more than four palmar and four dorsal interossei. The human specializations parallel those readily interpretable trends already noted in certain of the digits of other Primates, while the separate identities of the insertion tendons of flexor brevis and dorsal interosseous components are emphasized, not only by their typical mode of crossing to different insertions, but also by their relation to a tough, clearly defined transverse lamina. It must be emphasized that the transverse lamina is not part of the metacarpophalangeal joint but, together with the common extensor tendon, constitutes an entirely separate mobile sleeve investing the dorsum and sides of the joint. Its movement is enhanced by large bursae intervening between this extensor apparatus and the joint capsule. Besides the well-known dorsal bursa (which may communicate with the joint cavity) there is frequently a large bursa deep to the trans- PRIMATE MM. INTEROSSEI verse lamina on each side. Where only a palmar interosseous inserts the lateral bursa separates the transverse lamina from the collateral ligament of the joint; where a dorsal interosseous inserts the bursa intervenes between the transverse lamina and the subjacent phalangeal tendon, which in turn is usually separated from the collateral ligament of the joint by a second deep bursa. The presence of these lateral bursae (apparently not hitherto described) facilitates the clear and unequivocal demonstration of anatomical arrangements in this region. In the following account the number of hands, of the 25 dissected, showing any particular variant is indicated by the figures in brackets following the description. The hand of man shows, beyond reasonable doubt, the same seven flexores breves already observed in the other Primates; in addition, a first palmar interosseous (flexor brevis two) is usually separately identifiable. These flexores breves present a characteristic gross appearance, i.e. they are fusiform, have often a slender proximal tendon of origin from the metacarpal base but arise also from the metacarpal shaft, and certain of them show a marked tendency to merge with a related dorsal interosseous (cf. Colobus p e l y k m o s vellerosus). Flexores breves two, four, seven and nine, however, remain independent as the palmar interossei. Despite categorical statements to the contrary in the literature, many textbooks persist in ascribing to the palmar interossei a dual insertion, into both the extensor aponeurosis and the proximal phalangeal base. It is here confirmed that, with certain reservations, the normal insertion of a palmar interosseous is by a wing tendon into the extensor aponeurosis. To reach this attachment the tendon passes superficial to the transverse lamina, to which i t is firmly attached; an attachment which may be mistaken for a phalangeal insertion. In the case of the fourth palmar interosseous, however (as previously noted by Eyler and Markee, '54), the deep aspect of the wing tendon, just about the point where it is joined by the fourth lumbrical, often [ 111 becomes adherent to the underlying phalangeal base. This is clearly a secondary attach- 283 ment and wholly different from the direct primary insertion of the phalangeal tendon of a dorsal interosseous. (Occasionally there is a true phalangeal tendon on the radial side of the minimus, which will be considered with the fourth dorsal interosseous.) Theoretically, it might be expected that the first palmar interosseous would also retain an insertion as a wing tendon and, in fact, its tendon does often [ 111 pass completely superficial to a transverse lamina (here immediately dorsal to the metacarpo-phalangeal sesamoid) to reach the ulnar wing of the pollicial extensor apparatus. Such an insertion has been noted before by St. John Brooks (1886) and by Poirier and Charpy ('01). Most textbooks, however, give the insertion as the sesamoid and base of the proximal phalanx; in this present series the muscle was completely arrested at the sesamoid in only four cases but more often [ 8 ] was attached to sesamoid and wing jointly. Sometimes  the muscle is not distinguishable from the M. adductor obliquus which may then furnish the wing tendon. The descriptive human dorsal interossei are clearly composite muscles derived each from a dorsal interosseous proper (comparable to those of e.g. P a n ) together with a flexor brevis (numbers three, five, six and eight). The second dorsal interosseous exhibits this derivation in almost unmodified form, arrangements being virtually identical with those described for this region of the hand in P a n (fig. 2). There is variable fusion of the muscle bellies of the flexor brevis and the dorsal abductor (which may, however, remain separate) but their tendons retain complete individuality as they pass respectively superficial and deep to the transverse lamina to reach their appropriate insertions into the wing of the aponeurosis and the proximal phalangeal base. Similar features are usually shown by the fourth dorsal interosseous, where the component bellies are again varipbly fused and the flexor brevis is occasionally almost independent. The tendons show the usual insertions but part of the substance of the dorsal abductor may sometimes  accompany the flexor brevis tendon superficial to the transverse lamina and into the wing of the extensor aponeu- 284 0. J. LEWIS rosis, resulting in a diminution in the size of the phalangeal tendon. In one instance the whole of the dorsal abductor was thus inserted, there being no phalangeal tendon. In one further instance the dorsal abductor tendon passed as usual deep to the transverse lamina, but had lost its phalangeal insertion and gained instead an attachment to the deep aspect of the wing tendon. Such modifications, exceptional for the fourth dorsal interosseous, become the rule for the third. The component bellies of this third dorsal interosseous may occasionally be clearly differentiated but are most commonly inseparably fused : the tendons, however, again clearly indicate the muscle's phylogeny. Invariably a welldefined tendon passes superficial to the transverse lamina to join the wing of the aponeurosis. Rarely  a slender phalangeal tendon passes deep to the transverse lamina to insert upon the proximal phalanx. Sometimes  a tendon similarly passes deep to the transverse lamina but then joins the deep aspect of the wing tendon : in one such example such a tendon was separated from the transverse lamina by a bursa and retained a slender phalangeal insertion, emphasizing its derivation from the dorsal abductor component. Generally [ 171 however, the whole of the third dorsal interosseous inserts into the extensor aponeurosis by passing superficial to the transverse lamina. The constitution of the first dorsal interosseous is especially informative. It is a matter of commonplace dissecting room observation that a volar portion of this muscle, arising from the second metacarpal, wraps over the thick, phalangeal tendon of the remainder, in its passage to a more dorsal insertion. The dorsal part of the muscle, arising from first and second metacarpals and furnishing the phalangeal tendon, has long been known as the M. abductor indicis of Albinus (1734) and the volar portion as the M. interosseous prior indicis of Albinus (or M. flexor brevis indicis or M. extensor tertii internodii indicis of Douglas, 1750) (Henle, 1855; Wood, 1866, 1867, 1868; Le Double, 1897; Wood Jones, '49). This arrangement clearly recalls that seen in the comparable situation in Pun and in Colobus polyhomos vellerosus. Again the dorsal bipennate portion clearly represents the dorsal interosseous proper, and terminates as the usual thick phalangeal tendon gaining its insertion by passage deep to the transverse lamina, where the typical superficial and deep bursae may invest it. The volar portion (flexor brevis three) is variably separate and crosses the phalangeal tendon with a variant degree of obliquity, to reach the surface of the transverse lamina with which it merges. It may be prolonged [ l ] into the wing tendon (here mainly formed by the large first lumbrical), it may terminate as an abortive wing tendon adherent to the shaft of the proximal phalanx  but most frequently it simply merges into the transverse lamina . The common statement that the first dorsal interosseous inserts into bone only is clearly an oversimplification of the facts. The human dorsal interossei are thus generally comparable in arrangement with those of other Primates. One feature appears, however, to obscure their basic symmetrical pattern, the occasional occurrence  of a tendon on the radial side of the minimus, passing deep to the transverse lamina, whose muscle of origin attaches to the fifth metacarpal shaft between the origins of the ulnar head of the fourth dorsal interosseous and the fourth palmar interosseous, and which is, effectively, a delamination of the former. The tendon has the typical topographical relationships of the phalangeal tendon of a dorsal interosseous; it typically  inserts into the base of the proximal phalanx but may fail to reach this and so terminate on the metacarpo-phalangeal joint capsule . It is doubtless homologous with the fibrous band noted in the same situation in Pun, and with the common additional insertion into the fifth digit shown by the marsupial fourth dorsal interosseous; it is PRIMATE MM. INTEROSSEI probable that the bifid marsupial insertion of this muscle represents the primitive condition. CONCLUSIONS Clearly there obtains a basic plan of digital muscle architecture, equally applicable to man and to other Primates, which reaches its clearest expression when the transverse lamina is well developed, as in Pan and Homo. (The essentials of this arrangement are shown in fig. 2.) This basic pattern may be masked by partial fusion of the muscular bellies of certain of the flexores breves with those of the dorsal abductors with which they insert. Modifications of insertion also occur, but these specializations follow certain clearly defined trends, and have apparently occurred independently and repeatedly among the Primates. Thus, the prolongation of the flexor brevis into the wing tendon may be attenuated, leaving the muscle inserted into the transverse lamina; a dorsal interosseous phalangeal tendon, distal to its passage deep to the transverse lamina, may become incorporated with the deep aspect of the corresponding wing tendon, leading to diminution (or even abolition) of its primary phalangeal insertion; the dorsal abductor muscle may even merge completely with its related flexor brevis and the composite whole give rise only to a tendon passing superficial to the transverse lamina into the extensor aponeurosis wing. Examples of these various specializations have been described above in different Primates. It is thus apparent that the recorded numerical variations of the primate Mm. interossei depend on terminological, rather than morphological, considerations. The specializations noted above also clarify the mechanism by which the functional axis of the hand may become changed. Thus, if the belly of the second dorsal abductor (fig. 1 ) should fuse, not with flexor brevis five, but with the fourth muscle (whose tendon then comes to form its sole insertion) the palmar axis becomes shifted from the third to the second digit. A true 285 interpretation of the phylogenetic history of the human muscular arrangements can alone bring order and clarity to an otherwise complex and confused subject. The findings for the human hand in this study accord with Haines’ (’51) description of the general anatomy of the extensor apparatus and they confirm and extend Landsmeer’s (’55) findings regarding specific insertion details in the several fingers. Although strictly beyond the scope of the present account, it is noteworthy that arrangements on the ulnar side of the minimus accord with the basic finger pattern herein described. The superficial part of M. abductor digiti minimi inserts into the ulnar wing of the extensor aponeurosis; but, additionally, in every hand examined in the present series, a rounded and separate tendon was found to emerge from the core of the muscle (often from a distinctly separate belly) and to pass deep to the transverse lamina to a bony insertion into the base of the proximal phalanx. The M. abductor digiti minimi is thus fully comparable to any typical composite human dorsal interosseous (e.g. the second). This arrangement, apparently not previously described, seems to demand some revision of St. John Brooks’ (1886) classical account of the homologies of the hypothenar muscles. Acceptance of the view developed herein implies rejection of the alternative earlier and conflicting phylogenetic theories (fig. l ) , the foundations of which warrant scrutiny anew. Canonical terminology recognizes only four representatives of the primitive flexores breves - the four palmar interossei - in the human hand, and no homologues of the flexores breves are traditionally described as inserting with the dorsal interossei. Cunningham ( 1882) therefore theorized that the flexores breves which formerly occupied these latter situations must have disappeared; as has been shown, closer morphological analysis of the human dorsal interossei shows this view to be unfounded. The view postulating dorsal migration of certain of the flexores breves to become 286 0. J. LEWIS dorsal interossei was based solely on the interpretation of certain embryological findings, at a time when inflexible views on ontogenetic recapitulation prevailed. Ruge (1878) and Windle (1883) were thus seeking evidence of an ontogenetic migration of the future dorsal interossei, which they believed they had observed. Re-examination of their illustrations, however, makes it apparent that no actual migration occurs but instead a relative delay in the full development of the intermetacarpal parts of the dorsal interossei a delay associated with the later broadening of the hand and concomitant widening of the intermetacarpal spaces. McMurrich's ('03) view that two flexores breves are incorporated with each of the first and fourth dorsal interossei receives no support from human or comparative anatomy. Indeed, his figure of a section from the hand of a 6 cm human embryo appears amply to confirm the view herein proposed, namely, that each human dorsal interosseous results from the partial incorporation of a single flexor brevis with a dorsal abductor muscle, and that each palmar interosseous is a virtually unmodified flexor brevis. The observations made during this present study and the factual data recorded in the literature (although not necessarily their interpretation) lead in escapably to this conclusion. ACKNOWLEDGMENTS I should like to thank Professor A. J. E. Cave for providing the material used in this study and for valuable advice in the preparation of the manuscript. LITERATURE CITED Albinus, B. S. 1734 Historia Musculorum Ho. minis. Haak and Mulhovium, Leyden. Braithwaite, F., G. D. Channell, F. T. Moore and J. Whillis 1948 The applied anatomy of the lumbrical and interosseous muscles of the hand. Guy's Hosp. Rep., 97: 185-195. Bunnell, S. 1942 Surgery of the intrinsic muscles of the hand other than those producing opposition of the thumb. J. Bone Jt. Surg., 24: 1-31. Campbell, B. 1939 The comparative anatomy of the dorsal interosseous muscles. Anat. Rec., 73: 115-125. Cunningham, D. J. 1878 The intrinsic muscles of the hand of the thylacine (Thylacinus cynocephalus), cuscus (Phalangista maculata) and phascogale (Phascogale calura). J. Anat. Lond., 12: 434-444. -1882 Report on some points i n the anatomy of the thylacine (Thylacinus cynocephalus), cuscus (Phalangista maculata) and phascogale (Phascogale calura) collected during the voyage of H.M.S. Challenger in the years 1873-1876; with a n account of the comparative anatomy of the intrinsic muscles and nerves of the mammalian pes. Challenger Reports, 5: 1-192. Douglas, J. 1750 Myographiae Comparatae. Kincaid and Crawfurd, Edinburgh. Eyler, D. L., and J. E. Markee 1954 The anatomy and function of the intrinsic musculature of the fingers. J. Bone Jt. Surg., 36A: 1-9. Forster, A. 1916 Die Mm. contrahentes und interossei manus in der Saugetierreihe und beim Menschen. Arch. Anat. Physiol., Lpz., 1916: 101-378. Haines, R. W. 1950 The flexor muscles of the forearm and hand in lizards and mammals. J. Anat. Lond., 84: 13-29. 1951 The extensor apparatus of the finger. J. Anat. Lond., 85: 251-259. Henle, J. 1855 Handbuch der Systematischen Anatomie des Menschen. Erster Band. F. Vieweg und Sohn, Braunschweig. Hepburn, D. 1892 The comparative anatomy of the muscles and nerves of the superior and inferior extremities of the anthropoid apes. Part I. J. Anat. Lond., 26: 149-185. Keith, A. 1948 Human Embryology and Morphology, sixth edition, Edward Arnold, London. Landsmeer, J. M. F. 1949 The anatomy of the dorsal aponeurosis of the human finger and its functional significance. Anat. Rec., 104: 31-44. 1955 Anatomical and functional investigations on the articulation of the human fingers. Acta Anat. Suppl. 24. Le Double, A. F. 1897 Trait6 des Variations du Syst&me musculaire le l'homme. Schleicher Frbres, Paris. McMurrich, J. P. 1903 The phylogeny of the palmar musculature. Am. J. Anat., 2: 463500. Poirier, P., and A. Charpy 1901 Trait6 d'Anatomie Humaine. Tome 11. Masson et Cie, Paris. Ruge, G. 1878 Entwicklungsvorgange a n der Muskulatur des menschlichen Fusses. Morph. Jb., 4 (Suppl.): 117-152. St. John Brooks, H. 1886 On the morphology of the intrinsic muscles of the little finger, with some observations on the ulnar head of the short flexor of the thumb. J. Anat. Lond., 20: 645-661. 1887 On the short muscles of the pollex and hallux of the anthropoid apes, with special reference to the opponens hallucis. J Anat. Lond., 22: 78-95. Salsbury, C. R. 1937 The interosseous muscles of the hand. J. Anat. Lond., 71: 395403. Sonntag, C. F. 1923 On the anatomy, physiology and pathology of the chimpanzee. Proc. zool. SOC.Lond., I: 323-429. Stack, H. G. 1962 Muscle function i n the fingers. J. Bone Jt. Surg., 44B: 899-909. PRIMATE MM. INTER0 S SEI 1963 A study of muscle function in the h g e r s . Ann. roy. Coll. Surg. Engl., 33: 307-322. Thane, G. D. 1892 Quain’s Elements of Anatomy. Ed. by. E. A. Schafer and G . D. Thane. Vol. 11, pt. 2, 10th ed. Longmans, Green and Co., London. Windle, B. C. A. 1883 On the embryology of the mammalian muscular system. No. I - the short muscles of the human hand. Trans. R. Irish Acad., 28: 211-240. Wood, J. 1866 Variations in human myology observed during the winter sesson of 1865-66 287 a t King’s College, London. Proc. roy. SOC., 15: 229-244. 1867 Variations i n human myology observed during the winter session of 1866-67 a t King’s College, London. Proc. roy. SOC., 15: 518-546. 1868 Variations in human myology observed during the winter session of 1867-68 a t King’s College, London. Proc. roy. SOC.,16: 483-525. Wood Jones, F. 1949 The Principles of Anatomy as Seen i n the Hand. BailliBre, Tindall and Cox, London.