SMITHSONIAN MISCELLANEOUS COLLECTIONSVOLUME 72, NUMBER 11
THE ECHINODERMS AS ABERRANT ARTHROPODS
BYAUSTIN H. CLARK
(Publication 2653)
CITY OF WASHINGTONPUBLISHED BY THE SMITHSONIAN INSTITUTIONJULY 20, 1921
BALTIMORE, MD., U. S. A.
THE ECHINODERMS AS ABERRANT ARTHROPODSBy AUSTIN H. CLARKCONTENTS PAGEPreface iThe dominant characteristics of the echinoderms 2The larvce of the echinoderms 2The change from mid-somatic to inter-somatic development in the crinoids. 2The nervous system of the echinoderms 3The echinoderm coelome 5The water vascular system 6The echinoderm vascular system 7The echinoderm skeleton 8Autotomy in the crinoids 9The appendages of the crinoids gCharacters of a h}'pothetical crinoid with mid-somatic development only. 10The crinoids and the barnacles .10The crinoids and the aberrant barnacles 11The crinoids and the starfishes 14The sea-urchins and the brittle-stars 17The feeding habits of the echinoderms and of the crustaceans 17Professor Patten's interpretation of the affinities of the echinoderms.... 18PREFACEOf all the larger animal groups there is none which has been thesubject of such diversified opinion regarding its affinities and sys-tematic position as the Echinodermata. Originally assumed to berelated to the coelenterates on account of their radial symmetry, theechinoderms were later placed near the flat-worms, the annelids andthe chordates. At present they are regarded as representing a verydistinct and isolated group, some considering them as allied to thechordates, while others emphasize their points of similarity with theannelids ; most zoologists, however, are reluctant to commit them-selves regarding their probable affinities.For many years I have been convinced that the echinoderms areby no means such anomalous creatures as they appear to be, and that,in spite of their many and extraordinary peculiarities, they are un-doubtedly closely allied to the crustaceans, and especiJilly to the bar-nacles. In the following pages the reasons for this belief are given.Smithsonian Miscellaneous Collections, Vol. 72, No. 11
2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 'J2THE DOMINANT CHARACTERISTICS OF THE ECHINODERMSThe echinoderms are very anomalous forms, and their relationshipsto other animals are masked not only by a highly perfected radialsymmetry, but also by a unique development of all the organs of thebody. Their outstanding features are the presence of a vascular, arespiratory, and a superficial skeletal system, the last composed ofarticulated (calcareous) elements, the absence of gill clefts, and thesharp division of the body externally into (five radial) segments.In these features they agree only with the arthropods.THE LARV-'E OF THE ECHINODERMSThe unique larvae of the echinoderms, which are of very variedtypes, vermiform, bean-shaped with five ciliated rings and a long an-terior tuft of cilia, auricularise, bipinnariae, brachiolariae, plutei, etc.,aiid dift'er more or less widely among themselves in the details of theirdevelopment, are always at first bilaterally symmetrical, which maybe accepted as an indication that the echinoderms are derived frombilaterally symmetrical ancestors.The larvae do not grow directly into the adults, but the latter forthe most part arise from new growth within the larval body, thestructures peculiar to the larvre being absorbed ; in a few cases thedevelopment is direct.There is little in the structure or in the development of the echino-derm larvae which is comparable to the structure or to the develop-ment of the larvae of any other animals, and it is evident that theextraordinary features exhibited by the adult echinoderms have beenprojected so far forward in the ontogeny as quite to destroy thevalue of the larvae as phylogenetic indices.THE CHANGE FROM MID-SOMATIC TO INTER-SOMATIC DE-VELOPMENT IN THE CRINOIDSPerhaps the most interesting feature connected with the mor-phology of the crinoids, and one which it is necessary especially toemphasize in order to understand the relationships between them andthe other echinoderms, is the abrupt change in the regions of bodilygrowth and extension which takes place beginning with the formationof the arms. It is this sudden change from interradial to radial, orfrom mid-somatic to inter-somatic, development which occurs atthe commencement of arm formation that has always proved thechief stumbling block in the way of a correct interpretation of theseanimals.
NO. II ECHINODERMS AS ABERRANT ARTHROPODS CLARK 3In the echinoderms the primary appendages, the teeth, the waterpores, and the genital openings, and in the crinoids the primarynerve cords as well, are all interradial in position. There can be nodoubt, therefore, that the dividing lines between the five half somitesfall in the middle of the so-called radial areas, and in the echinoidsand asteroids a sharp line of division is always maintained down themiddle of the radial series of plates throughout life, while no suchdividing line is found in the interradial regions.In the young crinoids each of the somatic regions is completelywalled in by two large superposed plates, a dorsal basal and a " ven-tral "' oral ; beneath the basals there are usually 3-5 small infrabasalsalternating Avith them and corresponding with the oculars of theechinoids which, since they are entirely absent in large groups, andare usually more closely associated with the column than with thecalyx, are probably to be interpreted as a dissociated columnal.The young crinoid therefore has its body protected by ten largesomatic shields, five dorsal and five ventral, the latter with the primi-tive appendage under the median line.The arms first appear as evaginations in the intersomatic lines atthe plane of separation between the dorsal and ventral plates. Theevidence is that the skeleton of the arms is double, half being derivedfrom the somite on either side ; but whatever may be the ultimategenesis of their skeleton, the arms arise as linear and almost im-mediately biramous appendages taking their origin from the inter-somatic planes.From this point onward the development of the animal is entirelyintersomatic ; the peristome and its underlying nerve ring, the watertube ring about the mouth, the blood vascular ring, the genital cord,and the coelomic cavities all send ofif radial branches which increasein length as the arm grows, while along either side of the peristomialextensions (ambulacral grooves) there is formed progressively a con-tinuous series of reduplications (ambulacral lappets) of the littleflaps with their associated tentacles, the latter in communication withthe radial water tube, which surrounded the mouth in the " pre-brachial '' stage.THE NERVOUS SYSTEM OF THE ECHINODERMSIn the chordates the central nervous system never becomes sep-arated by mesodermal tissues from the tract of ectoderm from whichit originated in the embryo. Sedgwick remarks that this is a featureof all echinoderms in so far as the ventral nervous system is con-
4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. JZ
cerned, and when this nervous system is removed from- the surfacethe removal is efifected by invagination.But the only nervous system found in the echinoderms which inits details is at all comparable to the central nervous system of theother higher invertebrates is the so-called apical nervous system ofthe crinoids, which first forms relatively late in life, and which ap-pears to arise in connection with the coelomic epithelium.It seems to me that this nervous system of the crinoids, which ispossibly (though not by any means probably) represented by the so-called mesodermal nerve plexus in the starfishes, but which is quiteunrepresented in the other echinoderms, afifords the best indicationof the probable affinities of these animals, and at the same time itshigh state of development suggests that the crinoids have departedless widely from the ancestral type than have the other classes.In the developing crinoid the ectoderm of the surface of the bodybecomes more or less disintegrated, and its cells to a greater or lesserextent pass inward and intermingle with the cells of the underlyingmesoderm so completely that they can in no way be distinguishedfrom them, the body wall being formed of an ectoderm-mesodermcomplex in which the cells of the two types cannot be differentiated.On the inside of this body wall, apparently in connection with thecoelomic epithelium, the apical nervous system arises ; but in viewof the mixed nature of the component cells of this wall it seems notillogical to assume that this apical nervous system is in realityformed from cells which, originally ectodermal, have infiltratedthrough the underlying mesoderm and now appear as if they belongto the coelomic epithelium. If this hypothesis can be accepted it isobvious that the apical nervous system of the crinoids is in no waycomparable to the nervous system of the chordates, and as this is theonly echinodermal nervous system comparable with the nervoussystem of other animals it naturally follows that no affinity with thechordates can be inferred from the nervous system of the otherechinoderms.In the crinoids the larvae become attached by the ventral side of theanterior end, and the column is a development of the preoral lobe, aris-ing therefore from the place where in the larvae the anterior nervemass, just in front of the mouth, is situated. The mouth moves fromthe ventral surface onto the left side and then migrates upward,away from the point of attachment, until it comes to lie at the poleopposite the latter, that is, at the posterior end near the anus.
NO. II ECHINODERMS AS ABERRANT ARTHROPODS CLARK 5The larval nerves disappear; but some time after the torsion of theanimal is completed a new nerve mass forms in exactly the placethat the preceding nerve mass occupied, now become the point justabove the top of the column. It is thus natural to assume that thecentral mass of the apical nervous system in the adult crinoid, ap-pearing in the region previously occupied by the anterior nerve massof the larval crinoid, corresponds to it, and therefore that it repre-sents the anterior nerve mass in other invertebrates. But as a resultof the torsion which.the animal has undergone the right ccelomic sachas become extended anteriorly and, following the enteric wall, hasreached over onto the left side ; from its anterior end it gives ofif fiveanteriorly directed diverticula which at a later stage become cut ofifand give rise to the chambered organ about which the central nervemass forms a close and almost complete investment.The central nerve mass gives of¥ five stout cords which immediatelybranch, the two branches from each being connected by one or twocommissures, and then joining the similar branches from the nervetrunks on either side to form the radial nerves of the division seriesand arms. In each segment of the arms and pinnules, the nerve cordgives off from a ganglionic swelling four branches, two dorsolateraland two ventrolateral. In addition to these nerve cords, which maybe either single or double, and are in some types represented by twowidely separated parallel cords, each cirrus in its central canal con-tains a prolongation from the chambered organ ensheathed in nervefibers continuous with those of the central nervous mass, and fivesimilar prolongations from the chambered organ and the centralnerve mass extend downward into the stem.THE ECHINODERM CCELOMEMuch has been made of the fact that in the echinoderms thecoelome is enterocoelic in origin, as in the Brachiopoda, Chsetognatha,Chordata, and probably the Phoronida, while in the developing mol-luscs, annelids and arthropods the coelome is not enterocoelic in origin.Sedgwick notes that in all the chordates except the tunicates thecoelome in its first state in the embryo shows more or less markedtraces of three divisions, the anterior or proboscis coelome. which inthe Vertebrata and Enteropneusta is single and in Amphioxus double
;
the collar or middle coelome, which is always double, and the trunkcoelome, which is double and which in the Vertebrata and in Am-phioxus becomes metamerically segmented. In the echinodermsthere seem to be indications, at least, of a similar tripartite division;
6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 72there is the anterior coelome, which is sometimes single {Asterina)and sometimes double {Echinus), the hydrocoele, which is probablyfundamentally double, though in some cases • (holothurians and cri-noids) only one hydrocoele sac is formed, and the posterior coelome,which is alv/ays paired. But in those chordates in which the entero-coelic origin of the coelome is clearly presented these three divisionsof it always come off from the enteron separately, while in theechinoderms the enteron at most gives off only one pair of coelomicsacs; and whereas in the chordates the middle (collar) coelome isnever more closely associated with the anterior than with the posterior,in the echinoderms it is always closely associated with the anteriorcoelome, being developed from it and remaining connected with it bythe stone canal throughout life.Whatever its origin the coelome is clearly homologous in all thesetypes so that the manner of its development may be considered as dueto special mechanical or other limitations imposed by conditions inthe early stages—size. form, amount of yolk. etc.—and not to phy-logenetic causes.THE WATER VASCULAR SYSTEMThe most extraordinary structure of the echinoderm body is thewater vascular system. This arises as a narrow dorsolateral out-growth from a portion of the coelome which unites with an ecto-dermic infolding on the anterior aboral surface. From this developthe stone canal and the madreporites. The ectodermic opening placesthe hydrocoele in communication with the exterior, so that the organhas often been compared, in whole or in part, to an annelid excretoryorgan or nephridium.But Professor Patten has pointed out that it is much more likeone of the typical excretory organs of the arthropods (shell gland,green gland, coxal gland) which consist of thin walled coelomic sacswith a thick walled tubular outgrowth of varying length united toa short duct infolded from the ectoderm.The five primary tentacles or tube feet of the echinoderm larvaaccording to Professor Patten represent five modified thoracic ap-pendages ; an outgrowth of the underlying somite grows into eachappendage in typical arthropod fashion, but instead of breaking upinto separate muscles for the appendage it remains permanently inthe form of a membranous diverticulum of the hydrocoele and becomesthe distal end of a radiating water vascular canal. Only the distalend of the original appendage separates from the body as the primary
NO. II ECHINODERMS AS ABERRANT ARTHROPODS CLARK 7
tentacle ; the remainder of the appendage, however long it may event-ually become, may be regarded as lying in the surface ectoderm, de-veloping on either side as it increases in length paired cirri whichbecome the double row of tube feet for each arm into each of whicha prolongation of the water vascular canal extends.Primarily, then, each of the five primitive thoracic appendages, onefrom each of the five half metameres of which the echinoderm bodyis composed, contains a tube of nephridial intent leading into a canalopening to the exterior by a pore. The anterior and posterior endsof the series of half metameres join, and the excretory canal becomesa ring canal from which grow out five long radial canals giving ofifbranches to the tentacles or tube feet as these are formed.The excretory function of the water vascular system of the echino-derms is reduced to a minimum if, indeed, it can be said to exist at all.Its action is chiefly that of an hydraulic system whereby power origin-ating in a great number of weak and dissociated muscle fibers scat-tered along and within the water tubes and their branches is unifiedand transmitted to the hollow tube feet, tentacles, and other appen-dages, while at the same time the constant inflow and outflow ofwater through the madreporic openings, especially when these com-municate with the body cavity as they do in the gill-less crinoids,undoubtedly serve to a greater or lesser extent the purpose ofrespiration. THE ECHINODERM VASCULAR SYSTEMThe vascular system, which is especially well developed in theholothurians and echinoids, is formed of a peculiarly modified con-nective tissue in which the fibers are sparse and which contains inter-communicating spaces without an epithelial lining. The fluid in thesespaces does not appear to undergo any definite movement. Typicallythere is a circumoral tract with radial prolongations which lie be-tween the radial water vessel and the radial nerve cord, an annularaboral tract in which the generative rachis is embedded and whichsends oflf extensions to the genital organs, and in holothurians andechinoids a considerable development in the mesentery and on thegut wall.In the barnacles no heart is ever present, and the lacunar channelsin which the blood circulates are for the most part ill defined.
8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. J2.THE ECHINODERM SKELETONMacBride and others have remarked that the presence of calci-fied skeletal tissue in the mesoderm of the body wall is a characterfound in the echinoderms and vertebrates only among coelomateanimals. It does not seem to me that this can be considered as anyindication of affinity between these groups.Many vertebrates have uncalcified mesodermal skeletons, andothers have only partially calcified skeletons. The uncalcified skeletonof the notochord resembles in structure the parenchyma of the solidtentacles of certain coelenterates, and is quite different from anythingfound in the echinoderms. Calcareous deposits of greater or lesserextent occur in the mesoderm of barnacles, brachiopods, rotifers andcestodes, as well as in the mesogloea of sponges, and the calcareousskeletal structures of some coelenterates are mesodermal in origin.In certain of the early cystideans the skeleton appears to havebeen wholly or chiefly chitinous, and their surface exactly resemblesthat of the phyllopod and other crustaceans preserved in the same •rocks. It is not improbable, therefore, that the calcareous exoskeletonof the echinoderms of the present seas developed from a chitinousbody covering through an exoskeleton composed of chitin with anincreasingly greater amount of inorganic matter such as we see todayin most of the larger crustaceans.In the developing crinoid the ectoderm of the surface of the bodymore or less completely disintegrates and its component cells largelypass inward and intermingle with the cells of the underlying meso-derm, so that in the crinoid the outermost layer of the body is almostas much mesodermal as it is ectodermal. This, being the case, nomatter what its phylogenetic relationships and tendencies are, theformation of an ectodermal skeleton has now become impossible asthere is no continuous ectoderm from which to form it. A cal-careous mesodermal skeleton appears, the first rudiments of whichare formed in the deeper layers but soon move to a more superficialposition enclosing the body in a calcareous investment formed of largeand definite plates. Just before the appearance of the arms there are,in addition to the columnals, 13-15 thin cribriform films lying justbelow the surface and fitted edge to edge, including 3-5 infrabasals,5 basals alternating with them, and 5 orals superposed upon thelatter.Now although this skeleton is mesodermal and calcareous, the re-lations between it and the enclosed body of the animal are entirelydifferent from the relations between the vertebrate skeleton and the
NO. II ECHINODERMS AS ABERRANT ARTHROPODS CLARK 9body organs, being on the contrary quite like the relations betweenthe chitinous or more or less calcified skeleton of the arthropodsand their enclosed body. The crinoid skeleton is a superficial(though not external) skeleton enclosing the body and giving offarticulated appendages ; it therefore resembles the skeleton of thearthropods more closely than it does that of any other animals. Thefact that it is mesodermal and calcareous seems to me to be, in viewof its development in every way like an ectodermal chitinous exo-skeleton, and especially in view of the fact that it lies outside of theventral nervous system which runs along or just within its innersurface, of purely secondary significance.Crinoids are undoubtedly descended from animals with an ar-ticulated exoskeleton, and their articulated superficial skeleton iscalcareous instead of chitinous as would be expected solely becauseof the disintegration of their ectoderm in the young stages.In the arthropods we find in the sessile barnacles the beginningsof a transition from a chitinous exoskeleton to a calcareous mesoder-mal superficial skeleton, and from the conditions in these animals itis not difficult to supply the connection between the crustacean andthe crinoid skeleton.AUTOTOMY IN THE CRINOIDSThe crinoids with more than ten arms increase the number of theirarms by breaking off the larval arms at the base, the stump formingan axillary from which two or more arms arise. This is primarilydue to the inability of the brachial skeleton, a rigid calcareous invest-ment of the dorsal and dorsolateral portions of the arm, to keep pacewith the other brachial structures in development, and is thereforedistantly comparable to the moulting so characteristic of thearthropods. THE APPENDAGES OF THE CRINOIDSThe appendages in the crinoids are of two kinds. From the base,and always in connection with the chambered organ and the centralnen'e mass of the apical nervous system, arise uniserial jointed ap-pendages ending in a strong hook. These are especially developedin crinoids unprovided with a stem, and serve both as tactile organsand for attachment. In their position as anterior organs and intheir function as tactile and grasping organs, as well as in their uni-serial structure, thev recall the antennse of the barnacles.
lO SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. ^2About the ventral surface there are five long arms typically forkingat the base, thus representing biramous ventrolateral appendages andcalling to mind the biramous thoracic appendages of the barnacles.There can be no question of any direct homology between the cirriof crinoids and the antennae of barnacles, or between the thoracicappendages of barnacles and the arms of crinoids, but both sets oforgans have the same functions and the same location in each group,and are constructed on a similar plan, so that it is not impossible toregard them as parallel manifestations of the same ancestral ap-pendicular plan, a plan not occurring in the animal kingdom outsideof the arthropods and crinoids.CHARACTERS OF A HYPOTHETICAL CRINOID WITH MID-SOMATIC DEVELOPMENT ONLYLet us imagine a crinoid with entirely mid-somatic developmentand with entire instead of half somites, that is, with bilateral insteadof radial symmetry. We would have a body composed of five broadsomites each covered with a broad arched plate (tergum) to the edgeof which is articulated a flap (oral, corresponding to a pleuron) ;within this would be the radial (epimeron) at the inner edge ofwhich arises a biramous appendage. A body with five terga and fivepairs of biramous appendages with their bases covered by pleurafrom which they are separated by epimera would certainly be con-sidered as crustacean in character, and if it were attached by the headend, with the mouth upward, it would unhesitatingly be pronounceda barnacle, its deficiencies and anomalies of organization being as-cribed to degeneration.THE CRINOIDS AND THE BARNACLESThe crinoid develops from a highly anomalous larva, with a so-called vestibule suggesting a partial development of a bivalved cover-ing, which attaches itself by the anterior end like the cypris larva ofa barnacle and turns a half somersault bringing its mouth upwardand opposite the point of attachment, also like a barnacle ; so far itsdevelopment equally well suggests that of a polyzoan from a cypho-nautes larva ; but in its further growth it develops a superficialskeleton as does an arthropod of a sort already seen in rudimentaryform in the barnacles, with the chief nerve cords, which are highlydeveloped and arthropod-like, running over or just within its internalsurface as in the arthropods, and uniramous anterior (oriented fromthe central nerve mass) and biramous ventrolateral appendages,
NO. II ECHINODERMS AS ABERRANT ARTHROPODS CLARK IIboth sets of which have functions similar to those of the correspond-ing appendages in the barnacles.Thus in spite of the utter dissimilarity in the details the broaderfeatures of the structure of the crinoids and the barnacles as well asof their development are seen to be similar, or at least comparable,or perhaps it should be said that there is less divergence betweencrinoids and barnacles than there is between crinoids and any otherorganisms.THE CRINOIDS AND THE ABERRANT BARNACLESThe crinoids represent a derivative from a branch of the samearthropod stock that gave rise to the barnacles, but they have gonemuch further; half of each of the five segments of which the bodyis composed fails to develop so that the body is composed of fivehalf segments joined in a circle with the central organ of the ventralnervous system at one pole and the mouth, which has moved poster-iorly and come to lie near the anus, at the other ; the developmentof the body structures after early youth suddenly becomes entirelyinter-somatic instead of mid-somatic ; the peristomal region becomesenormously enlarged and extended, resulting in the formation of asort of lophophore ; and the appendages have been suppressed, or atleast appear in a very modified form late in life. Certain hereditarytendencies show themselves after the animal has, so to speak, re-covered from the profound ontogenetic shock resulting from theloss of half its body, in the appearance of articulated uniserial tactileand grasping organs at the neural pole (the original anterior end)and of articulated biramous appendages used for food gathering orfor locomotion along the ventrolateral border.Crinoids, like barnacles, are sessile, pedunculated, attached byhook-like processes, or unattached.The strong probability that the arthropod stock which by profoundmodifications gave rise to the barnacles also gave rise to the crinoidsis indicated not only by the asymmetry in the Verrucidae, in whichthe operculum consists of the scutum and tergum of one side only,those of the other side being fused to form one half of the wall whichis completed, on the side of the movable opercular plates, by thegreatly developed and displaced rostrum and carina, but also by theanomalous parasitic forms which have developed among the formerin which the aberrant features are so fundamental that they havebeen thrust forward into the ontogeny so far as to modify profoundlythe form and structure of the naupHus. These forms also show that
12 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 'J2the occurrence of complicated ramifying roots, so highly developedin some crinoids, is an inherent possibility in the barnacles.A combination of the asymmetry of the Verrucidse (inherent alsoin very many other crustaceans, and especially noticeable in thePaguridas and Bopyridae) carried to its logical conclusion in the com-plete atrophy of one side, with the modifications of the body seen inSphcsrothylacMs or Sarcotaces in a less extreme form, the roots ofthe Rhizocephala, and a skeleton formed after the manner of theplates in the shell of the Operculata, furnishes all the elementsneeded for recombination to form the crinoid. It may be well tocall attention to the fact that outside the Cirripedia there is no groupin which all the morphological peculiarities exhibited in the crinoidscoexist—indeed they are not to be found in all the rest of the animalkingdom together.In this connection it may be worth while to review the salientfeatures of the more important of the aberrant barnacles.The Rhizocephala are exclusively parasitic barnacles, the mostdegenerate of all parasites; in the adult stage they are distinguishedfrom normal barnacles by the entire absence of all traces of segmen-tation and of appendages, and at all stages they lack an alimentarycanal ; every trace of arthropod organization has disappeared. Nearlyall of them occur on decapod crustaceans.The body has the form of a simple sac, or may be divided intonumerous similar sacs, attached by a short peduncle from whichroot-like processes ramify throughout the body of the host ; theseabsorptive roots appear to be absent in the aberrant genus Duplorbis.The body proper is completely enveloped by the mantle which usuallyhas a narrow aperture capable of being closed by a sphincter muscle ;in Sylon the opening is double, and in Thompsonia, Clistosaccus andDuplorbis the mantle cavity is completely closed.The mantle commonly is attached to the visceral mass by a narrowmesentery near which on either side are the paired (more rarelyunpaired) openings of the male and female genital organs. In thedifferent genera the external form varies considerably, and with itthe position of the mesentery and of the genital apertures.Thompsonia, the most aberrant and highly specialized of all theparasitic barnacles, consists of nothing but a diffuse system of branch-ing and sometimes anastomozing mycelium-like roots continuousthroughout the body of the host and all arising from a single originallarva ; the peripheral division of the root system passes out into thewalking legs, abdominal swimmerets and tail fans and there givesrise to numerous (up to more than 500) small sacs consisting of a
NO. II ECHINODERMS AS ABERRANT ARTHROPODS CLARK 1
3
thin mantle without muscles and containing only an ovary, withoutgenerative ducts, testis or nerve ganglion. During development thevisceral mass disintegrates so that at the time of hatching the mantlecontains a great number of cypris larvae ready to emerge. Theescape of the larvae is contemporaneous with, or soon followed by,a moult of the host. The empty shells of the external sacs are carriedaway with the cast skin, and the terminal swellings of the root systememerge as a new crop of external sacs.In Peltogaster the body has an elongated sausage shape, with themantle opening at one end, and is attached by the peduncle about themiddle of its length. The mesentery is longitudinal on the proximalside (next the peduncle).In Sacculina the w^hole body is flattened in the plane of the mesen-tery and has assumed a secondary and superficial bilateral symmetryabout a plane at right angles to this and coinciding with the medianplane of the host. In other genera, such as Lerncoodisciis and Tri-angulus, the symmetry becomes still more complicated.In Clistosaccus and in Sylon the genital organs are impaired.The peduncle perforates the integument of the host and gives offon the inside the absorption roots which, in the case of Sacculina,penetrate into all the organs of the host with the exception of thegills and heart, and extend to the terminal segments of the legs andinto the antennules and eye stalks. In Duplorbis, in which the rootsystem appears to be absent, the peduncle is hollow, its cavity com-municating with the closed mantle cavity and opening at the otherend into the body cavity (h?emoccele) of the host.Apart from a single nervous ganglion (absent in Thompsonia)which lies close to the mesentery near the female genital openings,the only organs present are those of the generative system.SphcBTOthylacus is parasitic on a simple ascidian (Polycarpa), liv-ing attached by ramifying roots to the inner wall of the branchialsac. The globular body is enclosed in a mantle which has a smallopening. There are no appendages, but there is a complete alimen-tary canal with mouth and anus, the latter near the mantle opening.The two known species of Sarcotaces live embedded in the musclesof fish ; an alimentary canal is said to be present, and there are noroots.In the Ascothoracica, all of which are parasitic in Zoantharia orin echinoderms, the mantle may have a bivalved form (Synagogaand Petrarca), or it may form a capacious sac (Laura) much largerthan the body with which it is connected by a narrow neck and hav-ing only a small opening to the exterior. In Dendrogastcr the mantle
14 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. "J
2
is Still more developed and is produced into branched lobes. InLaura the mantle is covered with stellate papillae penetrating thetissues of the host and presumably absorptive. In all cases the mantlecontains ramifications of the enteric diverticula and portions of thegonads. In Laura the body is divided into six " thoracic " and threelimbless " abdominal " somites, and ends in a caudal furca. InPetrarca and in Dendrogaster the body is unsegmented.In these three genera a pair of preoral appendages is present and,except in Laura, are armed with hooked spines suggesting that theyare organs of fixation. They are inserted, at least in Laura, at thesides of the buccal region, and are more or less enveloped by themantle.The cement glands appear to be absent, and the mouth parts aremore or less reduced, but appear to be adapted for piercing.The thoracic appendages are biramous and articulated only inSynagoga. In Laura they are uniramous and indistinctly segmented,and the first pair are long and slender. In Petrarca they are stillfurther reduced, and in Dendrogaster they are represented only bysome indistinct papillae.In all three genera the gut ends blindly, and the hepatic diverticula,which are large, extend into the mantle. The nervous system isreduced. An eye is said to be present in Synagoga. In Laura theoviducts open at the base of the first pair of cirri.The larva of Laura is a nauplius lacking the frontolateral horns;in Dendrogaster the larva hatches as a peculiar cypris with only fivepairs of biramous thoracic limbs.In the Apoda the curious Proteolepas hivincta is elongated andmaggot-like, with no trace of a mantle nor of appendages other thana pair of adhering antennules. The mouth parts, borne on the first
" segment," seem to be adapted for piercing and sucking. The ali-mentary canal is greatly reduced ; according to Darwin only theoesophagus is present, and there is no trace of stomach, rectum oranus. THE CRINOIDS AND THE STARFISHESIn the crinoids during their development the mouth moves fromthe ventral surface onto the left side, indenting the left hydroceleand the left posterior coelome, and continues its migration until itcomes to lie at the posterior end beside the anus. When it hasreached this point the hydrocoele ring closes. As a result of thismovement the right posterior coelome has also shifted and come to
NO. II ECHINODERMS AS ABERRANT ARTHROPODS CLARK 1
5
lie on the aboral (originally anterior) side of the gut and the preorallobe of the larva becomes enclosed by the rows of skeletal elements(apical plates) which are developed outside the right posteriorcoelome in all echinoderms except holothurians ; these plates are firstlaid down in a horse-shoe shaped ring which later closes, as doesthe hydroccele, to form a complete ring.In the asteroids the closure of this curved row of plates is effectedfar from the point of origin of the preoral lobe on the right or rightdorsal side of the larval body ; in the crinoids it is efifected at theanterior end of the larval (posterior pole of the adult) body andencloses the preoral lobe just as the hydroccele does in asteroids.The larvae of both crinoids and asteroids attach themselves bythe preoral lobe ; but whereas in the crinoids the preoral lobe is quitefree of the circumoral vessel and arises from the apical or aboralsurface of the adult, in the asteroids the preoral lobe is encircledby the water vascular ring and its withered vestige springs from theoral surface of the adult disc.In the asteroids the mouth has shifted from the ventral surface ontothe left side of the body, but has gone no further. Since the preorallobe disappears as an appendage from the oral surface within thehydroccele ring it is evident that the apical nervous system of thecrinoids, which appears to be intimately connected with the preorallobe, must be represented by a ventral nervous system in theasteroids.In the crinoids the circumoral nerve ring and its extensions be-neath the ciliated ambulacral grooves of the disc, arms and pinnulesis associated only with the latter and with the ventral surface of thetentacles. In the asteroids the similarly situated ectodermal ner-vous tracts are in connection with a diffuse ectoneural plexus foundthroughout the ectoderm and at the mouth with an endoneural plexuswhich is the central portion of the so-called endodermal nervoussystem. The deep oral nervous system, consisting of a double cordin each radius just within the radial nerve thickening of the ectoneu-ral system and centering in a more or less complete ring about themouth, said to be exclusively motor in function, possibly correspondsto the deep oral system of the crinoids which consists of paired cords,one on either side of the water tube ; but the latter is connected withthe apical nervous system and the former is not. The apical nervoussystem of the asteroids, motor in function, consists of a cord in themid-radial line of each arm and appears to develop from the dorsalperitoneum with which it remains in continuity. It dififers fromthat of the crinoids in being radial (intersomatic) instead of inter-
l6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 'J2
radial (mid-somatic) in position, in consisting of single cords in-stead of paired cords with commissures, and in lacking a well de-veloped central organ. From its lack of any relation to the preorallobe and its general indefiniteness of structure it is difficult to see howit can be in any way homologous with the apical nervous systemof the crinoids. It is probably a special feature peculiar to asteroids.The apical nervous system of the crinoids, from which the wholeanimal with the exception of the peristomal region and its extensionsis innervated, is absent in the asteroids, or rather it has become mergedinto the circumoral structures and their derivatives.It has been shown that the crinoids before the development of thearms are encased in plates developed over the somatic divisions, butthat beginning with the appearance of the arms the development be-comes wholly intersomatic or radial. In the asteroids the develop-ment is radial from the first appearance of the plates, the interradial(mid-somatic) body covering seen in the young crinoids not appear-ing at all.In the crinoids there are at first no plates belonging to the ventralsurface, the basals being dorsal and the orals ventrolateral, but thelatter become ventral plates after the formation of the arms, whichthemselves are composed of a series of dorsal ossicles carrying exten-sions from the ventral structures on their ventral surface. In theasteroids the plates at their first appearance represent those of thecrinoids after the formation of the arms as far as their bifurcation,minus the orals. There is a central plate, corresponding to the infra-basals in the crinoids ; about this are five basals, corresponding to thefive basals of the crinoids ; beyond and alternating with these are fiveterminals, corresponding to the radials of the crinoids, but alwayssingle and showing no indications of a primarily paired condition as thecrinoid radials do. On the opposite (ventral) surface are five pairs ofplates, one pair in each radial division, corresponding to the first twopost-radial plates in the crinoids, but side by side instead of tandem.Thus whereas the extensions of the crinoid body—the arms—arejust on the border between the dorsal and ventral surfaces and arecomposed dorsally of dorsal ossicles and ventrally of extensions ofventral structures, in the asteroids the dorsoventral edge of the bodyhas moved dorsally so that the dividing line between the dorsal andventral surfaces falls between the radials and the succeeding plates,and the whole ventral surface is encased in plates which are repre-sented in the crinoids on the dorsal surface only.The asteroids therefore differ from the crinoids in the temporaryattachment of the larvce; in the relatively slight alteration in the posi-
NO. II ECHINODERMS AS ABERRANT ARTHROPODS CLARK IJtion of the mouth; in lacking the apical nervous system, which haseither completely disappeared or has become transformed into aventral nervous system centering in a ring or rings about the mouth
;
in omitting the early mid-somatic development, the development of thebody being inter-somatic from the first; and in having moved theborder between the ventral and dorsal surfaces dorsally so that allthe post-radial plates now lie on the ventral surface, the arms beingformed by an extension of the body in the plane dividing the crinoidradials from the plates succeeding.THE SEA-URCHINS AND THE BRITTLE-STARSThe echinoids and ophiurans differ still more widely from thecrinoid type. Their larvae, except in special cases, are extraordinarilydeveloped plutei which have no attached stage at all, and are char-acterized by the small size of the preoral lobe, by the great developmentof the post-anal portion of the body, and by the possession of a speciallarval skeleton supporting the arms which is later resorbed.Whereas the asteroids differ from the crinoids in transferring thepost-radial plates from the dorsal to the ventral surface and therebyforming a ventral skeleton of primarily dorsal elements, the echinoidshave gone further and have eliminated the dorsal surface altogetherexcept for a ring of plates about the periproctal region and the smallarea within it, the globular body being composed of plates representingthe ventral plates of the asteroids. The perfection of an entirely newtype of compact radially symmetrical body from the crinoid throughthe asteroid, simulating the compact radially symmetrical coelenteratebody, has furnished a starting point for new development, and bilateralsymmetry, superposed upon the perfected radial symmetry, has reap-peared and in some cases (as in Pourtalesia) has been carried to anextreme.The ophiurans are phylogenetically parallel to the echinoids, buttheir line of specialization is entirely different. In them the relationof the dorsal to the ventral surface has remained as in the asteroids,but the radial (intersomatic) extensions of the body have becomenarrowed and consolidated into highly efficient jointed appendagesfrom which all non-essential structures have been eliminated.THE FEEDING HABITS OF THE ECHINODERMS AND THECRUSTACEANSCorresponding with the progressive specialization in their structureit is interesting to note a progressive specialization in the feedinghabits of the echinoderms. The crinoids are plankton feeders, like the
l8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. ^2barnacles. The starfishes are largely carnivorous, feeding especiallyupon molluscs, but some swallow mud out of which they digest theorganic matter. The echinoids feed upon vegetable matter or uponorganic detritus, and many are mud or sand swallowers. Theophiurans feed largely upon detritus or swallow mud, but many areectoparasitic upon coelenterates and crinoids from which they stealthe food in or on its way to the stomach.Thus along with the widening of the gap between the structure ofthe true barnacles and that of the echinoderms there is a similardivergence in their feeding habits ; from plankton feeders they becomesimply scavengers and parasites.The recent crustaceans as a whole show exactly the same line ofspecialization from the phyllopod, which feeds on minute organisms,to the decapod, which feeds largely on carrion or detritus, or isectoparasitic on coelenterates or on crinoids, and this development ofthe feeding habit, with the emphasis on the scavenging activities, ischaracteristic of these two groups alone.PROFESSOR PATTEN'S INTERPRETATION OF THE AFFINITIESOF THE ECHINODERMSAlthough I arrived at the conclusion that the echinoderms and thearthropods are in reality closely related more than ten years ago andin a short paper published in April, 1910, that conclusion is readilyto be inferred. Professor William Patten was the first to attempt toexplain the relationships of the echinoderms to the arthropods indetail. His reasoning is so entirely dififerent from mine that it isworth while to repeat his arguments here.He says : " The echinoderms are notable for their contrasts and con-tradictions. Their outward appearance and their pronounced radialstructure distinguish them from all other animals, and at firstsight suggest a very primitive organization similar to that of thecoelenterates. On the other hand they display a high degree of his-tological and anatomical specialization that is in marked contrast withtheir low grade of organic efficiency. They begin their early em-bryonic development with a bilaterally symmetrical body and with clearindications of metamerism, only to change it in the later stages for onethat is radially symmetrical and in which all outward traces ofmetamerism have disappeared. After a short free-swimming larvalexistence they attach themselves, neural side down, by means of larvalappendages and a cephalic outgrowth ; they then turn neural side upand remain so attached for life; or in some cases they give up their
NO. II ECHINODERMS AS ABERRANT ARTHROPODS CLARK I9
sessile existence and again become free, moving slowly about, neuralside down. There are, therefore, three chief characteristics of theechinoderms that demand our first consideration: (i) The earlybilateral symmetry and metamerism; (2) the sessile life and modeof attachment b}^ cephalic outgrowths; and (3) the asymmetry.There appears to be but one explanation for these remarkable condi-tions, which is as follows : The early development of bilateral sym-metry and metamerism in the echinoderms, and the presence of atelocele and telopore in place of the more primitive gastrula andblastopore, clearly indicate that they had their origin in bilaterallysymmetrical animals of the acraniate type that had already acquireda considerable degree of complexity. These ancestral forms prob-ably belonged to the cirriped group, for before the latent asymmetrybecomes effective the young echinoderm larva resembles a cirriped inits form, mode of attachment, and subsequent metamorphosis morethan it does any other animal. The radiate structure of the laterstages was due to a persistent local defect, or to the absence of adefinite part of the embryonic formative material, which in turncreated a condition of unstable equilibrium, the result of which is thatthe whole side, following the path of least resistence, bends towardthe defective area, forming an arch that increases in curvature until anapproximate equilibrium is again attained by the union of the twoends to form a circle. The original half metameres and segmentalorgans are then arranged in radiating lines, thus creating a newradiate type and a new set of internal conditions that dominate thefuture growth of the organism. If we assume that a strongly markedasymmetry, such as that which occurs so frequently as an abnormalityin Xiphosura, or even as a normal character in the Bopyridse andPaguridse, was a fixed feature of the hypothetical ancestral cirripedsand was capable of a successful organic adjustment, we shall have aperfectly simple and natural explanation of the origin and structureof the echinoderms.
" The young asteroid larva is said to attach itself voluntarily atfirst, and for a short time only ; later it becomes permanently attached,head first and neural side down, in the same remarkable manner as ayoung cirriped, both the cephalic appendages (which are thick walledand muscular, with a long basal portion and a short terminal knobstudded with small adhesive papilhe, greatly resembling the minuteadhesive antennfe of the cirripeds and parasitic crustaceans) and theadhesive disc taking part in the process. The young crinoid larvaattaches itself wholly by means of the cephalic disc, as the adhesive
20 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 72
appendages appear to be absent. Its first position is with the neuralor oral surface down, as in the cypris stage of the cirriped. The discthen elongates, forming a slender cephalic stalk or peduncle, and thelarva turns a somersault, bringing its neural side uppermost. Mean-while the vestibule, or peribranchial chamber, which at first is smalland temporarily closed, enlarges, then ruptures, and the five ap-pendages project from the cup-like head in typical cirriped fashion.In certain of the representatives of the recent echinoderms, such asthe asteroids, the fixed stage is temporary, while in certain others,such as the echinoids and holothurians, it appears to be omitted alto-gether and the young echinoderm, after its metamorphosis, againacquires a limited power of locomotion. But in most primitive echino-derms, such as the stalked crinoids, blastoids and cystideans, a perma-nent attachment by an elongated cephalic stalk, in typical cirripedfashion, was the almost invariable rule, and no doubt represented theprimitive condition for the whole class. When an echinoderm doesbecome free it acquires only a very limited power of locomotion and ofcoordinated movement. Its characteristic lack of efficiency in thisrespect is due not so much to its simple structure as to the fact thatits freedom was gained at a late period in the phylogeny of a veryancient group in which sessile inaction was the prevailing condition.It is often assumed that a sessile or parasitic mode of life is the initialcause of degeneration. The various anatomical peculiarities commonto the copepods, cirripeds and acraniates do not bear out this con-clusion. The fact that in these diverse subphyla we see the same shift-ing of cephalic appendages to the haemal side, the same cephalic out-growths, and the same degeneration of the neuro-muscular organs,indicates that there are certain initial defects or peculiarities ofgerminal material common to the whole group, and that these are theunderlying cause of defective organization, the defective organizationbeing in every case of such a nature that a sessile or parasitic orvegetative mode of life is the only one possible."