..

•

-_ .

301

FIGURE 60.-Large and small bundles of fibers of the tensilium of C. tirginica seen in unstained and nondecalcifiedpreparation of the material teased in glycerin. Photomicrograph.

from circular to elliptical light areas as the planeof section of the fibrils becomes tangential (secfig. 62).

Sections made at right tLngles to the fibrils (fig.62) demonstrate a certain similarity to those ofthe organic membra.nes of the antgonitic part ofthe shells of mollusks and pearls. According toGregoire, Duchiiteau, and Florkin (1950, 1955),

54

such organic membranes have a lace-like structureconsisting of meshes and holes of different size andpattern. In these investigations by Belgianbiologists the material was first decalcified, andthe ],tyers of organic substance then separated byultrasonic oscillation to obtain the ultrathin mem­branes suitable for electron microscopy. Thefilms of the calcite-ostracum layer of the shells of

FISH AND WILDLIFE SERVICE

I<'IGURE 61.-Elcctron micrograph of the ligamcllt of C. virginica sectioned parallel to thc fibrils.

pelecypods which Illwe no true nacre (0. edulis,O. tulipa, Yoldia, Acra, and others) were found toconsist Hof heterogenolls mat.erinl, the morc repre­sentati\'e elements of which are amorpholls, \'itre­ous plat.es, somet.imes granular ll.nd de\-oid ofvisible (or unquestionable) pores." (l950, p.

THE LIGAMENT

30).' In the absence of ultrasonic equipment III

my Illboratory this method could not be used atWoods Hole, :\1ass. Comparison of figures pub­lished by Gregoire and his associates with thephotograph reproduced in figure 62 suggest.s t.hat

$ Translation by Paul S. Galtsoff.

55

",

... • , ,0 0.5

Microns ~FIGURE 62.-Electron micrograph of a section of the ligament of C. virginica made across the fibrils.

the structure of the ligament of G. virginica hassome similarity to that of the organic membranesof the aragonite shells. Recently Stenzel (J 962)has found that the resilium of the Ostreidae con­tains aragonite.

One of the sections of the ligamen t of C. virginicastudied with the electron microscope shows a seriesof black, oval-shaped bodies arrangcd along curvedlines and separated from one another by fibrils(fig. 63). The black bodies probably correspondto the small globules visible under the light micro­scope. Their nature has not been determined.

The action of the ligament can be demonstratedby a ratber crude model consisting of two sligbtlycurved pieces of wood, representing the valves,joined by a series of brass rods. The rods arehent and arranged to correspond to the COurse ofthe arches as the latter are seen in an enlargedphotograph of a transverse section of the ligamen t,(fig. 57). Thin rubber tubing interwoven betweenthe arches corresponds to the bundles of fibrils.Since the diameter of rubber tubing used in tbeconstruction of the model greatly exceeds thecomparable diameter of the fibrils, this portion ofthe model is not in scale. Another departure

56

from actual conditions is the interweaving of therubber tuhing between the arches, a method usedto simplify construction although no such arrange­ment of fibrils was disclosed by microscopy. Themodel is shown in fig. 64. If the sides of thestructure are pressed togetber, the arches curveup and exert lateral pressure at the same timethat the increased rigidity of tbe rubber tubingadds to the elastic force. One can easily feel thispressure by touching the rubber tubing with thefinger tips while bringing the "valves" together.

CHEMICAL COMPOSITION

The chemical composition of the ligament isessentially the same as that of the organic matrixof the shell (Mitchell, 1935: Trueman, 1949,1951). The proteins forming the lateral (tensil­ium) and the central (resilium) portions of theligament are not, however, identical. The differ­ence can be demonstrated by staining reactionsand by various chemical tests. For instance, inTellina tenuis the lateral parts of the ligament arestained red or yellow by Mallory triple stain,while the inner part turns blue, a differencecomparable to that between the staining reaction

FISH AND WILDLIFE SERVICE

,

•

Microns

FlllURE 63.-Electron micrograph of the ligament made ncar one of the arches parallel to the fibrils of C. virginica. Darkbodies probably correspond to the smallest globules seen in the light microscope.

of the conchiolin of the prismatic layer and thatof the calcite-ostracum discussed on p. 42. True­man (1949) concludes that the two types ofconchiolin seem to correspond respectively to thetwo components of the ligament. The tensiliumgives a positive reaction with the xanthoproteic,l\'lillon's, and Merker's reagents, whereas thereaction of the resilium to these reagents isnegative. Brown (1949) points out that most

THE LIGAMENT

of the epithelial skeletal proteins of invertehratesthat have been examined seem to be collagensand that their physical properties depend upondegree of hydration. The electron micrographsof the ligament (figs. 61 and 63) do not, however,show the axial periodicity of about 640 angstrom(A.) which is the most common characteristic ofcollagen fibrils (Gross, 1956). Other authorsdescribe fibrils of 270 A. period which participate

57

TABLE 10.-Results of chemical tests of the ligament ofTellina tenuis, according to Trueman

+

-Falnt. __

Test

Five percent HCI. _Saturated KOH (hot) _Xanthoproteic reactlon _Millon's reagent _Biuret reactlon _Nlnhydrln _Morner's reagent • _Chltosan test (Campbell) • _Chitin test (Schulze} _Argentaffine _

stance, the amber coloration of the lateral layer ofthe ligament is considered to be the result of tan­ning by an orthoquinone. This conclusion isbased on the fact that even after boiling this layerinduces rapid oxidation of the mixture of dimethyl­paraphenylenediamine and a-naphthol (Nadi re­agent), which is frequently employed to indicatethe presence of orthoquinones in the cuticles ofinsects and crustaceans (Dennell, 1947). In theligament of O. edulis the differentiation betweenthe two layers may be made visible by Mallorytriple stain. The lateral layer (tensilium) consistsof quinone tanned protein whereas the centrallayer (resilium) is built of calcified proteins (True­man, 1951).

Few chemical studies have been made on theligaments of oysters, but chemical analysis of thetwo portions of the ligament of the relatedpelecypod Tellina made by Trueman (1949) showsthe following differences summarized in table 10.

It is rather surprising to find that an elastic,nonliving structure functioning through a con­siderable period of time (according to Trueman,several years in Tellina) is heavily calcified. Theresilium of G. virginica contains a much largeramount of calcium carbonate than the outer parts:determinations made in my laboratory on theligaments of 5- and 6-year-old oysters dried at55° C. show that the calcium carbonate contentof the resilium varied from 30 to 67 percent of thetotal weight of the sample, while in the tensiliumthe content of calcium carbonate was only from5.3 to 8.5 percent.

It is apparent that knowledge of the chemistryof conchiolins and other substances found inmolluscan shells and ligaments is incomplete andthat much remains to be discovered about thecomposition and structure of these proteins whichplay such an important role in the life of allbivalves.

Outer I Innerlayer ~

----------------1---No effect

All dissolves++

t

FIGURE 64.-Mechanical model of the ligament of C.virginica. Arches are in scale and correspond to thecurves visible in a cross section of the ligament at amagnification of about 100 X. Diameter of rubber tUb­ing representing fibrillae is not in scale.

in the formation of the mature 640 A. periodcollagen (See pp. 512-513 of S. L. Palay [editor]Frontiers in Cytology, 1958), as well as smallerfibrils in the embryonic tissues. The latterprobably represent a very early stage in theformation of collagen.

Collagen fibers can be tanned in vitro, that is,they can be converted by various agents to a formin which they swell less and develop greaterchemical resistance. The tanning of proteinstructures by an orthoquinone occurs naturallyamong many invertebrates and has been demon­strated for the cuticles of a number of arthropods(Dennell, 1947; Pryor, 1940; Pryor, Russell, andTodd, 1946) and for the chaetae of earthworms(Dennell, 1949). There is also evidence that asimilar phenomenon takes place in the ligamentsof bivalves (Friza, 1932). In Anodonta, for in-

58 FISH AND WILDLIFE SERVICE

ELASTIC PROPERTIES

It has long been known that the ligament per­forms a mechanical function by automaticallypushing the valves apart when the tension of theadductor muscle relaxes. In a live oyster, how­ever, the gaping of the valves never attains thepotential maximum limited by the angle and lengthof the beaks. This can be demonstrated by asimple test: if the entire adductor muscle issevered, the valves open to a much greater anglethan that maintained by a fully narcotized oysterwith a completely relaxed muscle attached to theshell. It follows from this observation that duringthe entire life of the oyster the adductor muscle,even at the periods of its greatest relaxation, exertsa certain pulling force against the elastic tensionof the ligament.

In view of the voluminous literature dealingwith the structure and function of bivalve musclesit is surprising to find how little attention has beengiven to the study of the physical properties ofthe'ligament. The first attempt to determine thepulling force of the muscle sufficient to counteractthe elasticity of the ligament was made in a rathercrude manner in 1865 by Vaillant who tried tomeasure the elastic force of the ligament ofTridacna shells. Trueman (1949) erroneouslygives credit for this pioneer work to Marceau(1909), who only repeated the method used byearlier investigators (Plateau, 1884).

After removing the soft body of Tridacna,Vaillant set the empty shell on a table with theflat valve uppermost and placed a glass graduateon top of it. Water was poured into the graduateuntil the valves closed. Then the volume ofwater was read and its weight computed. Theweight of the water plus the weight of the glass con­tainer and of the valve gave Vaillant a value whichhe called the resistance of the ligament. For ashell of Tridacna, apparently one of small size, hegives the following figures: weight of water re­quired to close the valves-250 g.; weight of thevessel-700 g.; weight of the valve-632 g. Thetotal force needed to overcome "the resistance" ofthe ligament is, therefore, 1,582 g.

A similar method was used by'Plateau (1884),the only differences being that weights were addedto a metal pan suspended from a loop encirclingthe valves, as shown in figure 65, and that theshell was placed on a metal ring. The elasticforce exerted by the ligaments of several commonbivalves, as determined by Plateau, was found to

THE LIGAMENT

FIGURE 65.-Plateau's method of measuring elasticity ofthe ligament.

be as follows: 08trea eduli8-333.8 g.; Venusverruco8a-500.0 g.; Mya arenaria-620.0 g.; andMytilu8 eduli8-1,051.8 g. In Marceau's paperof 1909 the data taken from Plateau's work arerepeated without change or verification.

Trueman's investigation of the ligament ofTell1'na (1942) marks a renewal of interest in thestudy of the physical properties of the ligament.In a later paper (1951) he finds that in very youngO. edvli8 the outer layer of the ligament (ten­silium according to our terminology) forms acontinuous band along the entire dorsal marginof the hinge, but that in adults this outer layerseparates into the anterior and posterior portions,leaving the inner layer (resilium) exposed at thedorsal edge. The axis about which the valves ofthe adult O. eduli8 open (pivotal axis) is the samein C. virginica (figure 54, piv. ax.). In the closedshell of 08trea and Cra8808trea the central part ofthe ligament (the resilium) is under compressionand the two flanking portions (tensilium or outerlayer of Trueman) are under tension.

To measure the opening moment of thrust of ahinge ligament, Trueman (1951), uses the fol­lowing method, shown diagrammatically in figure66: soft parts of the body are removed and thelower valve embedded in plasticine; a counter­balanced beam is erected above the valve in sucha way that the weight placed on the pan at theleft end is applied at the center of the uppervalve. The distance from the left end of thebeam to the arm touching the centroid of the

59