Shell Layers and Structures in the BrackishWater Bivalve ...

Shell Layers and Structures in the BrackishWater Bivalve,

Corbicula japonica

Keiko Yamaguchi* +, a, Koji Seto* ,, Katsumi Takayasu* - and Morihiro Aizaki* +

Bivalve shells contain a great deal of information about the animal’s growth historyand local environment. Opaque and translucent layers are observed in the outer shellof Corbicula japonica. Detailed imaging with scanning electron microscopy and lasermicroscopy show that microstructural features are di#erent between the two layers.The di#erence is mainly determined by the content of organic materials ; the translu-cent layer contains less organic matrix than the opaque layer. A marking experimentrevealed that the translucent layers were formed in the period from early summer towinter, varying between individuals. If there is no formation of an opaque layer andshell growth rates are low in autumn, the translucent layer is accompanied by anannual ring on the shell surface. The period of formation of the layers is synchronizedwith the annual cycles of stable isotope values, especially with oxygen (d+2O). Bycombining the observation of the shell structure with stable isotope analysis, the shellof C. japonica can be used to provide information for reconstructing ecology andestuarine environments in the past.

Keywords : brackish water, bivalve, Corbicula japonica, microstructure, shell growth,translucent layer, stable isotopes

I. Introduction

Bivalve shells can contain a detailed recordof the animal’s growth history and environ-ment, similar to other accretionary growth skel-etons, such as corals or tree rings. The infor-mation stored within a bivalve’s hard parts isvery useful for the reconstruction of environ-mental changes. It is well known that bivalveshell chemistry holds a record of environmentalchange encountered by the animal. The car-bonate shell can provide a record of oxygenisotope variations (d+2O) that is often used aspaleo-thermometer (e.g. Craig, +30/ ; Rye andSommer, +32*). Growth-line analysis is alsorecognized as a useful method to investigatethe life history and growth rates of clams and

their response to environmental variation (e.g.Carter, +32* ; Dillon and Clark, +32*). Shellmicrogrowth patterns are useful for recon-structing tidal cycles or sea-level in the time andplace the animal lived (e.g. Lutz and Rhoads,+32* ; Ohno, +323). Compared to many other car-bonate skeletons (e.g. foraminifera, ostracods,nannofossils), the bivalve shell is very wellsuited for the reconstruction of detailed timeseries of environmental change covering thelife of the animal.

Shell also contains a great deal of informa-tion on the age and growth properties of theanimal. Age determination based on annualrings on the shell surface has been carried outin many marine bivalves (e.g. Orton, +3,0 ; Naka-oka, +33,). An annual ring is, in many cases, a

Received April 1, ,**0. Accepted June +3, ,**0. This paper was presented on the Symposium of the ,**/th Annual

Meeting of the Japan Assocoation for Quaternary Reaeach.

* +

* ,

* -

* a

Faculty of Life and Environmental Science, Shimane University. +*0* Nishikawatsu, Matsue, 03*�2/*., Japan.

Research Center for Coastal Lagoon Environments, Shimane University. +*0* Nishikawatsu, Matsue, 03*�2/*., Japan.

Shimane University. +*0* Nishikawatsu, Matsue, 03*�2/*., Japan.

Corresponding author : [email protected]. ac. jp

�� The Quaternary Research� ./ � / � p. -+1�--+ Oct. ,**0

slow-growth zone formed when the shell growthstopped in winter (winter break). The spawn-ing break in shell growth is also recognizedin many marine bivalves (e. g. Rhoads andPannella, +31* ; Sato, +33/).

Though marine bivalves have often been theobjects of such studies, applying these tech-niques to brackish water bivalves has not beencommon. As environmental variation associ-ated with climate change may be amplified atthe boundary between land and sea, studies ofenvironmental records stored in the shell geo-chemistry and microstructure of brackish bi-valves become more valuable.

Corbicula japonica is an endemic bivalve as-sociated with the low salinity brackish environ-ment. It inhabits brackish lakes and estuariesfrom Hokkaido to Kyushu in Japan, and is avery important species for inland fisheries inJapan (Nakamura, ,***). Paleo-environmentalinformation can be provided from the shells ofC. japonica, not only from fossil deposits butalso from shell middens associated with an-cient human habitation.

Kobayashi and Takayasu (+33/) studied shellmicrostructure of Japanese brackish-water cor-biculids in detail and observed that the outershell layer of C. japonica is composed of the com-plex crossed lamellar structure (finely-crossedtype). Takayasu et al. (+330) carried out a mark-ing experiment to examine the process of mi-crogrowth increments of C. japonica in LakeShinji, and found that the microgrowth patternshowed neither cyclicity nor periodicity. Theysuggested that the formation of microgrowthincrements was most likely influenced by com-plex changes in the brackish water environ-ment as well as by physiological rhythms andindividual variation in these physiological fac-tors.

On the contrary, several studies on shellgrowth and age determination of C. japonicawere carried out based on annual rings in shell(Utoh, +32+ ; Takada et al., ,**+ ; Oshima et al.,,**.). Although some specimens have clearlyrecognizable annual rings, it is usually di$cultto identify annual rings on the shell surface,owing to the thick, dark colored periostracumand the existence of irregular concentric ringscovering the surface of the shell. Utoh (+32+) ob-

served that the annual ring is associated withthe translucent layer within the mineralizedshell of C. japonica, collected from Lake Aba-shiri. He recognized that the shell of C. japonicahas two characteristic layers, translucent andopaque, and he suggested that the translucentlayer is formed when the shell growth rate islow. The translucent layer has been recognizedin other bivalves, e.g. Anadara (Kobayashi, +310),Gomphina (Kobayashi, +313), Mytilus (Hosomi,+323), and been regarded as an annual marker.Kobayashi (+310) reported that shell structureof the outer shell layer of Anadara broughtoniichanges between crossed lamellar and compos-ite prismatic structures alternately, and thatthe crossed lamellar structure is rather trans-lucent and formed accompanied with a slow-growth zone. Hosomi (+323) observed whiterings corresponding to the translucent layer inshells of Mytilus galloprovincialis, and used themfor age determination. But he did not describethe microstructure of the increment nor thetime of formation of the rings, and whether it isrelated to environmental factors or not. Prezantet al. (+322) investigated color change in fresh-water Corbiculid shell layers, and argued thatcolor change occurs when an animal is exposedto a significant stress. This change is betweenpurple and white layers in the shell. Kobayashiand Takayasu (+33/) also mentioned that thecalcareous shell layers of Corbiculidae speciesare composed of wider dark and narrower lightband arranged alternately. These probably cor-respond to the translucent and opaque layers,but there has been no detailed observation ordescription of these shell layers in the litera-ture.

Before shell layers can be used as the basis ofstudying variation in the brackish water envi-ronment, more work needs to be done on shellstructure and growth banding, particularly fac-tors controlling opaque and translucent layers.The purpose of this study is to describe thestructural di#erence between the opaque andtranslucent layers and to document the forma-tion of the two layers. Finally we will examinethe possibility of using these layers for the agedetermination, growth analysis, and paleo-envi-ronmental reconstruction.

Oct. ,**0Yamaguchi, K., Seto, K., Takayasu, K. and Aizaki, M.318

II. Materials and methods

+. MaterialsLiving individuals of Corbicula japonica were

collected between December +333 and June ,**/

from Lakes Shinji, Jinzai, and Nakaumi, easternShimane Prefecture, southwest Japan (Fig. +).Specimens used in this study were about , cmin length, typical of adult-size clams. They livedin shallow water along the coast of these brack-ish lakes. Because the tidal amplitude in theselakes is very small, clams remain covered bywater at all times.

The salinity of the water at the habitat of C.japonica in Lake Shinji is about / psu. That ofLake Jinzai is somewhat variable, between -�+2 psu. The salinity of the shallow water inLake Nakaumi is higher, +* to ,* psu. The wa-ter temperature ranged from , to -,� in all thelakes.

,. Observation of shell structureA single valve from each specimen was sec-

tioned from umbo to the ventral margin to ex-amine the shell structure. The sectioned valvewas observed with a binocular microscope, andthe thickness of the layer formed each monthwas measured. Then it was polished and etchedwith *.+N HCl for -* seconds or + minute. Thesurface was coated with Pt�Pd or Au by ion-sputtering. The section was then observedusing a binocular reflected-light microscope(Leica MZ�+*), a laser microscope (KEYENCE,V+/**) and a scanning electron microscope(SEM : JEOL JSM�/2** and HITACHI S�2**).Scanning electron micrographs in this paper

are secondary emission images (SEI) taken ataccelerating voltages of /4* or +* kV. A lasermicroscope (KEYENCE VK�2/**) was used toproduce a -D image of the etched shell section.

The outer shell layer of one sectioned valvewas sub-sampled at fine spatial resolution (*4.

mm thick layers) along the shell surface to ex-amine the change of oxygen and carbon stableisotope values throughout growth and in rela-tion to shell layers. A gas-ratio mass-spectro-meter (Finnigan MAT delta S) was used tomeasure the oxygen and carbon stable isotoperatios. Analyses were calibrated relative tothe PDB standard and the standard deviationof repeated internal standards was less than�*.*/�.

-. Field experimental methodologyA field experiment was conducted to analyze

the growth of the C. japonica and to test thetiming of formation of the translucent layer.

Two artificial ponds were used for culturingC. japonica beside Lake Jinzai. Brackish lakewater was continuously supplied to the ponds.Pond dimensions were +* m in length, / m inwidth and ,*�0* cm in depth.

Clams were distributed throughout one ofthe ponds at a density of ,4* kg/m, in wetweight. This density was determined at theconclusion of another unrelated study usingthese bivalves (Maeda et al., ,***). The averageinitial wet weight of the individuals was -4- gon 1th March ,***. This experiment was re-started on +*th April ,**+, and the initial wetweight of clams averaged ,42 g.

The clams were collected at random from the

Fig. + Index maps of Lake Shinji, Lake

Nakaumi and Lake Jinzai

,**0� +*� Shell layers in Corbicula japonica 319

artificial pond using three -*�-* cm quadratsevery month. The specimens were brought tothe laboratory and a number of growth relatedvariables were measured (wet weight, shelllength, shell height, shell width, wet flesh weight,dry flesh weight and ash weight). Mean animalsize was calculated from these values everymonth. The new generations entered into thepond after the start of the experiment wereexcluded from the average, as they could bedetected easily by the di#erence in shell size.

Condition index value (C.I.) was calculated torepresent the body mass status in the shell.Condition Index is defined as below (Suemitsuet al., ,**+) ;

C.I.�AFDW��SH�SL�SW�

AFDW : ash free dry weight of soft parts,SH : shell height, SL : shell length, SW :

shell widthTo examine the time of formation for trans-

lucent and opaque layers, .0 clams were la-beled with numbers and enclosed within a cagein the pond. These .0 individuals were num-bered on the outer shell surface (Fig. ,�A).These numbered individuals were cultured inthe cage for one year, from 1th March ,*** to ,

nd March ,**+. The position of the shell edgewas marked every month by peeling the peri-ostracum back slightly with small knife blade(Fig. ,). Shell height of each individual wasmeasured with a caliper at the time of edgemarking. The marked specimens were recol-lected from the cage after the field experimentand brought to the laboratory. After each shell

was cross-sectioned and polished, the period offormation of each shell layer was identified byreferring the shell height of each month andthe shell edge marking trace in the cross sec-tion.

Monthly growth rate was normalized as thelength of shell increment per day, calculatedfrom the increase of shell height and the num-ber of days of the experimental period.

III. Results

+. Occurrence of the translucent, opaqueand semi-translucent layers

Alternating translucent and opaque layerswere observed on the transverse shell sectionof Corbicula japonica using an optical binocularmicroscope (Fig. -). Opaque layers made up alarge majority of the shell.

The translucent layer is the bright layer inthe transmitted light image ; however it ap-pears as a dark layer in the reflected light im-age because it absorbs light (Fig. -). The opaquelayer has the opposite appearance, e.g. it isbright in the reflected light image as it reflectslight (Fig. -).

There is also a type of layer with intermediatecharacteristics, semi-translucent. This layer ismainly seen adjacent to the ventral side of theopaque layer. Because light is scattered withinboth the translucent and semi-translucent lay-ers, the boundary of these layers is not clearunder the binocular microscope.

,. Shell structure (SEM Observation)The outer shell layer of C. japonica is com-

posed of the complex crossed lamellar struc-

Fig. , The marking trace of the shell of Corbicula japonica

A) The sample identified as number “,”. Concentric white lines on the ventral side are the edge-marking

traces, B) Cross section of the ventral shell of A. Small arrows : the edge-mark traces.


ture (finely crossed type) as Kobayashi andTakayasu (+33/) reported. Both translucentand opaque layers are composed of this com-plex crossed lamellar structure. On a fine scale,however, the appearance of these two layers isvery di#erent (Figs. .�A�F).

In the scanning electron microscope image,the translucent layer appears as a white zone,while the opaque layer appears as a dark zone(Fig. .�G). Many micro growth striations witha width of a few micrometers can be seen ina dense pattern on the etched surface of theopaque layer (Figs. .�A, C). These striationsare grouped into distinct growth lines of +* to,* mm intervals (Fig. .�A). The opaque layer ischaracterized by many weak lines and a fewstrongly defined ones. In contrast only a fewstrong growth lines at about +* mm intervalspacing are recognized on the surface of thetranslucent layer (Fig. .�B). The strong growthlines in the translucent layer seem to be inter-ruptions of the crystal growth (Fig. .�D), whilethe weak striations in the opaque layer do notnecessarily represent breaks in microcrystal-lite growth (Fig. .�C).

At higher magnification, the microcrystallitesurfaces are also di#erent. In the opaque layerthe etched surface of the microcrystallites hasa smooth and rounded appearance (Figs. .�C,E), while that of the translucent layer has arough and angular appearance (Figs. .�D, F).In the opaque layer, the organic matrix whichoriginally enclosed the lamella but now re-mains behind after the etching of calcium car-bonate crystals, is present in the image andcauses the smooth surface (Fig. .�E). In con-trast, little organic matrix remains on the

etched surface of calcium carbonate crystals inthe translucent layer (Fig. .�F). Here the crys-tals are deeply etched and this results in arough and porous surface. This rough surfaceinduces edge e#ects in the SEM and the trans-lucent layer appears as a bright region in theSEM image (Fig. .�G).

Figure / shows a fractured section of the twolayers. The di#erence between the two layersis not clear in fractured sections, and any di#e-rence is subtle. The crystal cleavage surfaceappears clearer and sharper in the translucentlayer than in the opaque one. No porous struc-tures are seen in the fractured section of thetranslucent layer, which suggests that the po-rous texture of the etched sample is the resultof corrosion by the etching process (Fig. .�F).

To estimate the corrosion resistance of thetwo layers, a specimen was observed by thelaser microscope with the computer image pro-cessing system to check the surface topogra-phy. Figure 0 shows the boundary of the twolayers. The -D profiling image shows the highresistance of the opaque layer to corrosion (Fig.0�B).

-. Layer boundaries and the semi-translu-cent layer

At the transition from opaque to translucentlayers, the boundary is sometimes not clear(Figs. - and .�G). A reflected light image of thelatter shows that an intermediate transitionalzone exists between the two layers. This inter-mediate zone is a little more translucent thanthe opaque layer. This “semi-translucent layer”is also recognized by SEM observation (MD inFig. .�G), where the semi-translucent layer hasthe intermediate features between the opaque

Fig. - Transverse section of the shell of C. japonica observed by binocular microscopy


and translucent layers (Fig. .�H). The etchedsurface of semi-translucent layer is rougherthan that of the opaque layer, but the organic

matrix can be seen forming micro-growth lineslike the opaque layer.

.. Stable isotope chemistry

Fig. . Scanning electron micrographs of the transverse section of the shell of C. japonica

Arrows indicate “strong lines” among the striations. A) The opaque layer, B) The translucent layer, C) Higher

magnification of A, D) Higher magnification of B, E) Higher magnification of C, F) Higher magnification of D,

G) The boundary of the two layers. OP : opaque layer, MD : semi-translucent layer, TR : translucent layer, H)

The semi-translucent layer at the same magnification as C and D. Growth direction : downward.


Figure 1 shows the shell layers and the oxy-gen and carbon isotopic values. The position ofsample numbers (+�-.) is shown in Figure 1�C. The d+-C values ranged between �240 and�,4.�, and d +2O between �34/ and �.4-�.The cycles of d+-C and d+2O values within theshell are almost synchronized with each other.The pattern of isotope fluctuation goes througha repeated series of characteristic changes : [+]a relatively stable and slowly decreasing d+-Cand d+2O values (Sample number : --�,/, +2�+,, 2�/), [,] a region of rapidly decreasing val-ues (,/�,,, +,�++, /�.), and [-] a rapidly in-creasing slope (,,�+2, ++�2, -�+). There is atendency for the translucent layer to be formedwhen the isotopic value is increasing rapidly,while the opaque layer is primarily formedduring the gradually decreasing isotopic val-

ues. This pattern of change in the isotopicratios is especially apparent in d+2O, which isbest synchronized with the formation of theshell layer types. The length of one isotopiccycle on the shell section gets shorter as theshell grows.

/. Culturing experiment+) The timing of translucent layer formationThe annual survival rate of the C. japonica in

the artificial pond was about /*�. Twentyindividuals of the marked clams were recol-lected after the experiment and used to docu-ment the period of the shell layer formation.The growth period and the type of layers weredetermined by referring visible edge marks ina polished shell cross-section to the record ofshell height for each month. The shell layertype was evaluated under the binocular micro-

Fig. 0 Laser microscopic image of the shell section, polished and etched with HCl for one minute

A) Reflecting light image of the boundary of the translucent layer (left side) and the opaque

layer (right side). The translucent layer appears to be dark image as the light is absorbed.

B) Image of the same site of A, profiled -D topography. The surface level of the opaque

layer is higher than that of the translucent layer after its higher resistance to the corrosion.

Fig. / Fractured section of (A) opaque layer, (B) translucent layer


scope. As the precise boundaries of the semi-translucent layer were di$cult to identify, thesemi-translucent layer is included in the trans-lucent layer classification.

The timing of translucent layer formationwas di#erent among individuals (Fig. 2�C). Theformation of the translucent layer began inMay or June for some individuals, and the ma-jority had begun producing translucent layers

by late summer. Some individuals stoppedforming the translucent layer and began toform the opaque layer in autumn (Fig. 2�A).Other individuals formed the translucent orsemi-translucent layer in autumn, but the thick-ness of the translucent layer in autumn wasvery thin in these individuals, as discussedbelow. In the last cases, growth of the translu-cent layer was slow throughout the summer

Fig. 1 Cross section of C. japonica used for stable isotope analysis

The shell of C. japonica was collected in Lake Shinji (entrance of Ohashi River) in December +333.

A) Photograph using an optical microscope. White lines are the cutting markers, B) SEM image

of A. C) Location of sample numbers +�-., D) Corresponding variation of shell layer types and

isotopic values.


and continued into winter period, forming onetranslucent layer (Fig. 2�B). In some individu-

als, two translucent layers were formed duringone year, in May and autumn, but the layerformed in May was very thin.

,) Shell growth ratesSeasonal change of the average growth rate

is shown in Figure 3�A. The clams grow rap-idly from spring to early summer, but thegrowth is apparently inhibited in August. Whilea small amount of growth is recognized inautumn, growth stopped during the wintermonths. Shell height is reduced by a smallamount from December to January, resultingfrom shell corrosion during the winter dor-mancy.

Figure 3�B shows shell height of individualnumbered specimens. Almost all the individu-als increased in size from May to July. FromAugust to September, most individuals showedvery little growth, although some continued togrow. From October to November some showedmoderate growth, but others showed littlechange. Individuals that showed considerablegrowth in autumn produced the opaque layerduring that period, while those that showedonly little growth built translucent layer shell.

-) Condition index and dry tissue weight

Fig. 2 The marked period of shell formation in C. japonica

A) A specimen that produced an opaque layer in autumn, B) A specimen showing little growth and forming the

translucent layer in autumn, C) Shaded bar indicates the growth of translucent (including semi-translucent) layers

in every specimen examined. The specimen shown in A is ID number , in C, while B is number - in C.

Fig. 3 Seasonal growth in shell height of C. japonica

cultured in the artificial pond beside Lake Jinzai


The condition index (CI) represents the rela-tionship of soft tissue mass to valve size. Thesoft parts within the shell increased in massfrom spring to early summer, decreased through-out summer, and increased again from autumnto early winter (Fig. +*). In winter, CI decreasedslightly. The average trend in CI follows thispattern every year in both the artificial pondand in Lake Shinji.

0. The translucent layer in C. japonica inLake Nakaumi

Figure ++ shows a shell cross-section of C.japonica collected from Lake Nakaumi. Theopaque layer is clearly dominant. Althoughthe translucent layers are present, they arevery thin. The annual ring (representing aperiod of slow/no growth) is recognized ac-companied by the thin translucent layer.

IV. Discussion

+. Significance of the translucent layerThis study shows that the opaque and trans-

lucent layers of Corbicula japonica have di#e-rent microstructure characteristics as seen inscanning electron microscopy, although both

are composed of the crossed lamellar structure.The translucent layer is thought to include lessorganic matrix among the calcareous shell ma-terial than the opaque layer, based on the re-sponse to shell etching. The di#erence in cleav-age faces in the fractured sections indicatesthat the crystals of the translucent layer arelarger, more regular, and incorporate little or-ganic matrix. The lack of organic inclusionswithin the lamellar matrix probably resultsin the higher transparency of the translucentlayer.

It has been suggested that the modificationof shell structure serves a functional purposein some bivalves. The chalky layer developedwithin the shell of oysters, for example, is amodification of the foliated layer (Yamaguchi,+33. ; Chinzei, +33/). In this case, calcite foliagrow rapidly forming a box-work structure,resulting in a porous structure within the foli-ated layer. This box-work layer is very fragilebut allows for very rapid inflation of the shellform by the addition of fast-growing shell atlow metabolic costs. Chinzei (+33/) also pointedout that the chalky layer provides a functionaladvantage by contributing to a light-weightshell. In contrast, when a porous structure ispresent in the translucent layer of C. japonica(Fig. .�F), it is formed by the corrosion of theshell material and is not part of a functionalstrategy producing lightweight shell structures.The translucent layers can be seen in the shellof some photo-symbiotic bivalves (e.g. Ohno etal., +33/ ; Carter and Schneider, +331). Thesetransparent layers are used to keep the photo-symbiotic algae within the mantle of the bi-

Fig. +* Change of the condition index of C.japonica

cultured in the artificial pond of Lake Jinzai

Error bars : standard deviations.

Fig. ++ Shell section of C. japonica collected from Lake Nakaumi

A) Transverse section of the shell, B) Higher magnification of enclosed area in A.


valve (e.g. Fragum fragum). The translucentdomains of the shell consist of fibrous pris-matic structures in these species (Cardioidea).There are no reports of C. japonica keepingphoto-symbiotic algae, and therefore the trans-lucent layer most likely does not have a specialfunction, rather it is produced by some con-straint of shell construction, most probably thelow content of organic material.

Some studies have shown that translucentand opaque layers make an annual cyclicity(e.g. Kobayashi, +310 ; Utoh, +32+ ; Hosomi, +323).Kobayashi (+310) studied the structural changeof outer shell layer of A. broughtonii. He showedthat the composite prismatic structure (opaquelayer) was formed from spring to autumn,while the crossed lamellar structure (translu-cent layer) was formed in winter. In the case ofC. japonica, in this study, however, both thetranslucent and opaque layers are composed ofthe crossed lamellar structure. Thus the inter-mediate “semi-translucent” layer is available inthis case. Moreover, in the crossed lamellarstructure of C. japonica in this study, the trans-lucent layer is formed at variable times bydi#erent individuals and in both the warmperiod and cold season of the year. This indi-cates that the change of appearance in thecrossed lamellar structure does not depend onwater temperature. Rather the translucent layeris formed when the animal’s resources requirelimiting the metabolic cost of producing newshell material as discussing below.

,. Timing of the translucent layerThe marking experiment revealed that the

timing of translucent layer formation was vari-able among individuals ; although mainly inthe summer, it varied between May and winter.This indicates that the formation of the layerdepends on some internal physiological process,one that can vary between individuals. Mostlikely it is related to reproductive activities.The condition index shows the reduction ofsoft part mass occurs from June to October(Fig. +*). The decrease of the condition index ismost likely driven by reproduction. The spawn-ing period of C. japonica has been identified asfrom June to October in Lake Shinji by Saka-moto (+323), or as “mainly in summer” by Naka-mura (,***). This corresponds to the main pe-

riod of translucent layer growth. Spawningrepresents the largest demand on organic ma-terial production for clams. Thus the alloca-tion of organic material to shell matrix may bedecreased at the same time that the soft bodymass is reduced and the animal requires moreenergy and organic material to invest in repro-duction. A distinct spawning break in shellgrowth is observed in some bivalves, e.g. Mal-letis sp. (Rhoads and Pannella, +31*) and Phaco-soma japonicum (Sato, +33/), but is not clear inC. japonica. The spawning period of C. japonicais very long in warm lakes like Lake Shinji,compared to those bivalves, and therefore shelland gonad growth may both occur during asignificant portion of the year. We suggest thatthis leads to the production of a shell type thatis energetically more economical rather than acomplete cessation of shell growth.

In summer, moreover, the time of highestwater temperature (sometimes higher than -*�in these lakes) is associated with reproduction.Metabolic activity becomes higher as the watertemperature increases, but filtration rate is re-duced above ,/� (Nakamura et al., +322). Thisindicates that the clam is also stressed by highwater temperatures. This suggests that bothincreased reproductive demands and reducedfood intake occurs in mid summer, leading to alack of organic material availability. This isthe most likely reason that in August most ofthe individuals produce translucent layers ofthe shell (Fig. 1).

In some individuals the translucent layer isconstructed from autumn into winter. C. japon-ica burrows deep into the sediment when thewater temperature falls lower than /�+*� inwinter (Goshima et al., +333). The animal accu-mulates a nutritional reserve for the winterdormancy as indicated by the condition index(Fig. +*), which also results in the reduction ofthe allocation of organic material to shell ma-trix. The thickness of the translucent layerformed before the winter dormancy is verythin, because the rate of shell growth in theclam is very low at this time. The growth ofthe soft body in autumn mainly refills the shellcavity after the condition index has been re-duced during the summer.

In the case of adult C. japonica resident in


Lake Nakaumi, the translucent layer is verythin (Fig. ,). This is most likely a layer con-structed during the accumulation of nutritionreserves for the winter. The growth rest zone,produced by the winter growth-cessation, canbe seen just after the formation of the thintranslucent layer. In this specimen a translu-cent layer produced by spawning stress is notrecognized. Baba et al. (+333) indicated thatadequate temperature and salinity is impor-tant for C. japonica to produce a healthy repro-ductive condition of the gonads. The salinityof Lake Nakaumi is much higher than that ofLakes Shinji and Jinzai, and the high salinitymay have reduced the reproductive activity ofthis species and led to minimal production of atranslucent layer during spawning.

Utoh (+32+) thought that the translucent layeris formed when the shell growth rate is de-creased. This does seem to be the case in thetranslucent layer formed in autumn to winter,but the experiment in Lake Jinzai revealed thatthe translucent layer formed in June or July isproduced during a time of high growth rate.

Thus the formation of the translucent layeris not controlled by water temperature or growthrate. It is much more closely tied to times whenthe allocation of organic material to the shellmatrix is reduced.

-. Shell layers, age determination, and en-vironmental analysis

The winter break, recognized as a resting, orno-growth, zone in the shell surface is used forage determination in many bivalves (e.g. Naka-oka, +33,). Utoh (+32+) demonstrated that theannual ring in C. japonica from Lake Abashiriis a brown line on the shell surface and is cor-related with the translucent layer. He thus re-garded the translucent layer as the annual ringand used it for age determination. In manycases for shells of C. japonica in Lakes Shinjiand Jinzai, however, it is di$cult to identifyannual rings on the external shell surface, ow-ing to the dark colored and thick periostracumand to the existence of many irregular concen-tric rings. This makes the age determination ofthese clams di$cult. The marking experimentin Lake Jinzai showed that the amount of shellgrowth in autumn is di#erent among individu-als in this study. In the cases where autumn

growth is small, the translucent layer is formedfrom summer through the autumn and intowinter (Fig. 2�B). In these cases, the layer isbroad and continuous through the last half ofthe year, producing a marker of the annual layer.Shell growth of C. japonica in Lake Shinji, in-vestigated by Oshima et al. (,**.), increasedas the water temperature became higher, butstopped in August and did not resume in theautumn or winter. In these cases, the forma-tion of the translucent layer is limited to justbefore the winter growth break, and it can beused as the annual ring for judging the age ofthe clams. In a colder region, such as LakeAbashiri, shell formation of clams is limited tothe early summer season ; they grow only fromMay to July and show no growth in autumn(Utoh, +32+). Moreover their reproduction ac-tivity is limited to mid summer, from July toAugust. Thus shell formation is limited to theperiod before spawning. In Lake Abashiri wa-ter temperature declines rapidly in autumnand remains low from October to May. Thesefactors lead the annual rings to correspond to agrowth break beginning in late summer andstretching into the next spring. So the for-mation of the translucent layer always corre-sponds to the winter growth break in thesecases.

This is not always the case, however. Indi-viduals showing relatively high growth ratesin autumn produced opaque shell at that time(Figs. 2 and 3). In the artificial pond of LakeJinzai, two translucent shell layers can beformed within one year ; in summer and beforewinter. These two translucent layers are verydi#erent in appearance. The layer formed insummer is relatively thick because the animalcontinues shell growth during the allocation ofresources to reproduction. The layer formedjust prior to winter is very thin because shellformation is reduced when water temperaturebecomes lower. Thus the translucent layer canbe counted as an annual ring, provided thatthin translucent layers are not included in thenumber.

Oshima et al. (,**.) removed the periostra-cum before identifying and counting annualrings on the shell surface. This may be a usefulmethod to easily count annual translucent lay-


ers. A thick translucent layer creates a darkregion on the surface of the mineralized shell.The translucent layer seems to be a more easilyidentified annual landmark in the shell thanthe resting zone on the shell surface.

As the example of the shell grown in LakeNakaumi shows, reduced reproduction in moresaline waters results in the availability of moreorganic matter for shell growth and, just as ajuvenile before reproductive maturation, theanimal does not pay a large metabolic cost forreproduction and produces only a thin translu-cent layer prior to winter. This shows that care-ful consideration is required when shell layercharacteristics are used for the age determina-tion, especially for small shells.

The timing of each shell layer’s growth isclosely related to the variation in stable isotoperatios in the shell, especially the d+2O value ofthe shell layers (Fig. 1). The oxygen isotoperatio of carbonate shells has been used as apaleo-thermometer in the marine environment(e.g. Craig, +30/ ; Rye and Sommer, +32*). Inthis study we observed that the translucentlayers were constructed in the period from sum-mer to winter (isotope regions [,] summer to[-] autumn in the sequence discussed above).Moreover, the gradually declining isotope ra-tios (isotope region [+] in the sequence) withinthe opaque layers suggest that the shell growsrapidly from spring to early summer. Thesetrends are consistent with the interpretationdiscussed above. Thus in cases where there isuncertainty about whether the translucent layeris annual or not, stable isotope ratios will helpclarify the issue. Of course, if the data is avail-able, cycles in the d+2O values within the shellwill illustrate the growth and age of the clam,as can be seen in the decrease of annual growthrates shown in Figure 1�D.

The isotopic values, however, also varywith changing salinity (e.g. Ingram et al., +330 ;Dettman et al., ,**.). The collection location ofthis specimen was the Ohashi River, where thesalinity tends to change daily. The isotope pro-file from this specimen indicates that the influ-ence of temperature upon the stable isotopecomposition of the shell is greater than that ofsalinity. This process may be reinforced if theclam closes its valves or ceases its shell growth

when the salinity of the surrounding waterchanges rapidly due to estuarine conditions.On the other hand, the low d+2O value of sampleno. ,, and the high value of no. 2 may reflectlow salinity and high salinity events in the wa-ter derived from Lake Shinji. In order to dis-cuss the relationship between the water qual-ity and isotopic values in the shell layers, higherresolution sampling, using a micro-drilling meth-od, of shell layer material is required (Dettmanand Lohmann, +33/). The combination of ob-servation of the shell structure and stableisotope chemistry analysis, as shown by thisstudy, provides a powerful tool for reconstruct-ing or analyzing estuarine and coastal environ-ments.

Growth-line analysis is a useful method inthe study of the life history and growth of bi-valves and their environment. The pattern ofmicro-growth increments in samples of thisstudy shows neither cyclicity nor regular pe-riodicity, similar to the observations of Taka-yasu et al. (+330) in Lake Shiji. This study dem-onstrated that the opaque and translucent lay-ers have di#erent patterns in growth striationlines ; the growth striations in the opaque layerconsist of both strong (i.e. well pronounced andclearly visible) features (at +*�,* mm intervals)and numerous weak lines (at a few mm inter-vals), while those of the translucent layer areonly strong ones (at �+* mm intervals). Thedaily shell increment in May was over .* mm onaverage (Fig. 3�A). About +* weak striationscould be formed per day in the opaque shelllayers when the shell is growing rapidly. Judg-ing from these observations, the weak striationin the opaque layer is formed by a frequentchange of the environment at the shell growthsurface, such as opening and closing the valves.The strong growth lines seem to be growthbreaks in the construction of the shell, but thefactor controlling the periodicity of the breakis not clear because of the irregularity of theintervals. This is the subject of additional studyin near future.

Acknowledgement

We acknowledge Dr. David Dettman in Ari-zona University for improving the manuscript.We would like to thank Professor Ohshima of


Shimane University for allowed us to use theelectron microscope. The Shimane PrefectureFreshwater Fisheries Experimental Station alsohelped us to take SEM photographs. Ms No-hara in Shimane University helped us to takethe -D graphs with the Laser Microscope. Thisstudy is supported by the Grand-in Aid of theJapanese Society for the Promotion of Science(++02*/1-, +--*2*-, and +/,*.*.0).

References

Baba, K., Tada, M., Kawajiri, T. and Kuwahara, Y.(+333) E#ects of temperature and salinity on spawn-ing of the brackish bivalve Corbicula japonica inLake Abashiri, Hokkaido, Japan. Marine EcologyProgress Series, +2*, ,+-�,,+.

Carter, J.G. (+32*) Environmental and biological con-trols of bivalve shell mineralogy and microstruc-ture. Rhoads, D.C. and Lutz, R.A. (eds.) Skeletalgrowth of aquatic organisms : 03�++*, Plenum Press.

Carter, J.G. and Schneider, J.A. (+331) Condensinglenses and shell microstructure in Corculum (Mol-lusca : Bivalvia). Journal of Paleontology, 1+, /0�0+.

Chinzei, K. (+33/) Adaptive significance of the light-weight shell structure in soft bottom oysters.Neues Jahrbuch fur Geologie und Palaontologie.Abhandlungen, Abt.B, +3/, ,+1�,,1.

Craig, H. (+30/) The measurement of oxygen isotopepaleotemperatures. Tongiorgi, E. (ed.) Stable iso-topes in oceanographic studies and paleotempera-tures : +0+�+2,, Laboratorio di Geologia Nucleare.

Dettman, D.L., Flessa, K.W., Roopnarine, P.D., Schone,B.R. and Goodwin, D.H. (,**.) The use of oxygenisotope variation in shells of estuarine mollusks asa quantitative record of seasonal and annual Colo-rado River discharge. Geochimica et CosmochimicaActa, 02, +,/-�+,0-.

Dettman, D.L. and Lohmann, K.C. (+33/) Approachesto microsampling carbonates for stable isotope andminor element analysis : Physical separation of sam-ples on a ,* micrometer scale. Journal of Sedimen-tary Petrology, A0/, /00�/03.

Dillon, J.F. and Clark, G.R. (+32*) Growth-line analysisas a test for contemporaneity in populations. Rhoads,D.C. and Lutz, R.A. (eds.) Skeletal growth of aquaticorganisms : -3/�.+/, Plenum Press.

Goshima, S., Ikegawa, M., Sonoda, T. and Wada, S.(+333) Seasonal vertical migration within sedimentby the brackish clam Corbicula japonica. BenthosResearch, //, 21�31.

Hosomi, A. (+323) The ecology of Mytilus. +-1 p, San-kaido. (J)

Ingram, B.L., Ingle, J.C. and Conrad, M.E. (+330) Iso-topic records of pre-historic salinity and river in-flow in San Francisco Bay Estuary. Hollibaugh, J.T.and Center, R.T. (eds.) San Francisco Bay : The eco-system : -/�0+, American Association for the Ad-

vancement of Science.Kobayashi, I. (+310) Internal structure of the outer

shell layer of Anadara broughtonii (SCHRENCK).VENUS, -/ (,), 0-�1,. (J�E)

Kobayashi, I. (+313) Microtopography of the innershell surface of Gomphina melanaegis ROMER. Sci-ence Reports of Niitaga University, Series E, /, +*1�++3.

Kobayashi, I. and Takayasu, K. (+33/) Internal shellmicrostructure of two species of Japanese brackish-water. LAGUNA, ,, --�/+. (J�E)

Lutz, R. A. and Rhoads, D.C. (+32*) Growth patternswithin the molluscan shells. Rhoads, D.C. and Lutz,R.A. (eds.) Skeletal growth of aquatic organisms :,*-�,/., Plenum Press.

Maeda, I., Aizaki, M., Yamaguchi, K. and Fujita, N.(,***) A study on water purification by the Bivalve,Corbicula japonica, using outdoor experimentaltanks with continuous flow system. Journal of Ja-pan Society on Water Environment, ,-, 1+0�1,*. (J�E)

Nakamura, M. (,***) The Japanese fishery of Corbicula.,00 p, Tatarashobo Press. (J)

Nakamura, M., Yamamuro, M., Ishikawa, M. andNishimura, H. (+322) Role of the bivalve Corbiculajaponica in the nitrogen cycle in a mesohaline la-goon. Marine Biology, 33, -03�-1..

Nakaoka, M. (+33,) Age determination and growthanalysis based on external shell rings of the proto-branch bivalve Yoldia notabilis Yokoyama in O-tsuchi Bay, northeastern Japan. Benthos Research,.-, /-�00.

Ohno, T. (+323) Palaeotidal characteristics determinedby micro-growth patterns in bivalves. Palaeontol-ogy, -,, ,-1�,0-.

Ohno, T., Kato, T. and Yamasu, T. (+33/) The origin ofalgal-bivalve photo-symbiosis. Palaeontology, -2,+�,+.

Orton, J.H. (+3,0) On the rate of growth of Cardiumedule. Part I. Experimental observations. Journal ofthe Marine Biological Association of the UnitedKingdom, +., ,-3�,13.

Oshima, K., Suzuki, N., Nakamura, M. and Sakura-moto, K. (,**.) Shell growth and age determinationof the brackish water bivalve Corbicula japonica inLake Shinji, Japan. Fisheries Science, 1*, 0*+�0+*.

Prezant, R.S., Tiu, A. T. and Chalermwat, K. (+322)Shell microstructure and color changes in stressedCorbicula fluminea (Bivalvia : Corbiculidae). TheVeliger, -+, ,-0�,.-.

Rhoads, D.C. and Pannella, G. (+31*) The use of mol-luscan shell growth patterns in ecology and pale-oecology. Lethia, ., .+-�.,2.

Rye, D.M. and Sommer, M.A. (+32*) Reconstructingpaleotemperature and paleosalinity resimes withoxygen isotopes. Rhoads, D.C. and Lutz, R.A. (eds.)Skeletal growth of aquatic organisms : ,//�,3., Ple-num Press.

Sakamoto, I. (+323) Breeding season of the brackish


bivalve, Corbicula japonica in Lake Shinji. Zoologi-cal Science, 0, +,+3.

Sato, S. (+33/) Spawning periodicity and shell micro-growth pattern of the Venerid bivalve Phacosomajaponicum (Reeve, +2/*). The Veliger, -2 (+), 0+�1,.

Suemitsu, K., Yamaguchi, K. and Aizaki, M. (,**+)Study on the mechanism of the large amount deathof Corbicula japonica. LAGUNA, 2, -3�.0. (J�E)

Takada, Y., Sonoda, T., Nakamura, M. and Nakao, S.(,**+) Growth and settlement of the bivalve Cor-bicula japonica population in Lake Shinji. NipponSuisan Gakkaishi, 01, 012�020.

Takayasu, K., Kobayashi, I., Morita, H. and Naka-mura, M. (+330) Observation of the micro-growth

striations in the brackish-water clam, Corbicula ja-ponica from Lake Shinji. LAGUNA, -, +*-�++*. (J�E)

Utoh, H. (+32+) Growth of the brackish-water bivalve,Corbicula japonica PRIME, in Lake Abashiri. Sci-ence Reports of Hokkaido Fisheries ExperimentalStation, ,-, 0/�2+. (J�E)

Yamaguchi, K. (+33.) Shell structure and behaviourrelated to cementation in oysters. Marine Biology,++2, 23�+**.

(J) in Japanese, (J�E) in Japanese with English ab-stract.

��

��* +, a��* ,�� * -��* +

��

�� !�"#$#%&'()#*+,-. /01234�5�6��7896:896(;<=*-. >�?@A:BCDC?@AEFG%HIJKLMNO� P6�@GQR��9=S%T,(;<=*M. 7896UVW896�01XYZ[�\]^(_%,M<�� P6(`%-abJcd:ef=*M. ghij�U-:� 896�kS=l�

Sm+n�"*� o�pq�r��UL+"#$#Es-. t�7896(Euv� ��(_%,wx�� 896:��yz�{|(}~d-. �6�n�� :�q�+,-. ��QR:��Js��-�:�UV� /01234��J�L+�� !J��d-FG%&'J�-�:(Eu-:q�"*-.

�� : �� !�� "#� $%�

* +

* ,

* -

* a

�� ¡�^¢£¤¡¥ ¦03*�2/*. §¨©ª«¬ +*0*.

�� ¡��®¯°±²³C ¦03*�2/*. §¨©ª«¬ +*0*.

�� ¡ ¦03*�2/*. §¨©ª«¬ +*0*.

Corresponding author : [email protected]. ac. jp

,**0{ +*´ Shell layers in Corbicula japonica 331

Shell Layers and Structures in the BrackishWater Bivalve ...

Documents

Transcript of Shell Layers and Structures in the BrackishWater Bivalve ...