CHEMICAL COMPOSITION OF DREISSENA POLYMORPHA (PALLAS, 1771) SHELLS IN THE LAKE DĄBIE (WESTERN POMERANIA)

The paper presents results of derivatographic analysis applied to determine the basic chemical composition (organic matter and calcium carbonate) of Dreissena polymorpha (Pallas, 1771) shells. Derivatographic analysis was supplemented in the present study with assays of organic carbon (Corg), SiO2, Al, Fe, Mg, K, Pb, Co, Zn, Cu, Ni, Cd, Mn and S in homogenised samples. Key WordS: Bivalvia, Dreissenidae, shell composition, element content Folia Malacologica 5/1993 was originally published as No. 1462 of Scientific Bulletins of University of Mining and Metallurgy, Cracow. This digitalised version was prepared by the Association of Polish Malacologists and first published on-line on December 30th, 2016. http://dx.doi.org/10.12657/folmal.005.007 Folia Malacol. 5: 119–130

Dreissena polymorpha (PALLAS, 1771) is a bivalve common in Polish waters and is very important in aquatic communities due to its role in biofiltration and biosedimentation (e.g.STANCZY"KOWSKA 1968, 1977, WIKTOR 1969, PIESIK 1983).For these reasons, it is important to know the ,chemical composition ofthe shell.WIKTOR (1969) touched upon that problem in his paper.
Shells of bivalves and gastropods are important in formation of lacustrine carbonate sediments.Thus the knowledge on the basic chemical composition of shells, particularly in terms of thermal properties of organic matter, of as many species as possible, will contribute to elucidation ofthe origins of these sediments.
The paper was aimed at determining changes in chemical composition of Dreissena polymorpha shells in relation to their length, and at presenting statistical relationships between the chemical components and shell length as well as between various components themselves.An attempt was also made to relate shell chemistry to that of unconsolidated lacustrine sediments Dreissena populations dwell on.

Material and methods
Materials were collected, with a van Veen grab operated on board the SNB-1 motor boat, from the bottom of the Lake Da,bie in late July/early August 1988.
Well preserved shells (i.e.devoid of any sign of a mechanical damage) of dead bivalves were selected for the analysis.The shells were thoroughly rinsed in distilled water, fouling organisms and sediment remains being carefully removed.As the shells tended to break down in two halves, these were taken for the analyses.The clean shells were dried at 105°C to constant weight, which took about 4 h.Then the shells were divided into length classes, each class containing the same number of left and right halves; the same number of shells was ascribed to each length interval (0.1 mm).The samples were subsequently weighted to obtain the mean shell weight in a length class.The derivatographic analyses were performed in a 1500 D Derivatograph (MOM Budapest) with a simultaneous recording of differential thermal (DTA), thermogravimetric (TG), derivative thermogravimetric (DTG), and temperature (T) curves.The latter is an auxiliary curve used to determine temperatures of individual reactions and is not shown in the derivatogram enclosed.The derivatographic analyses were carried out under the following constant conditions: sample weight of 400 mg; curve sensitivity: 200 mg for TG, 500 V DTA, 250 V for DTG (500 V from 600°C up); heating rate of 10°C/min; atmospheric air; ceramic crucible; Al 2 0 3 as inert substance.
Organic carbon (Corg) was determined with Tiurin's method at the Szczecin University's Chemical Laboratory.
Assays to determine Si0 2 , Al, Fe, Mg, K, Ph, Co, Zn, Cu, Ni, Cd, Mn, and S contents were performed in an ICP Spectrophotometer (Baird, The Netherlands) at the Chemical Laboratory ofZD "Cuprum" KGHM Lubin.

Results
Due to insufficient amount of material available, Dreissena shells smaller than 4 mm were not studied.The derivatographic analysis can be run on any sample weight.However, it is purposeful to use equal sample weights in every analysis.The approach adopted in the present study was to analyse averaged samples consisting of a number of shells rather than analysing individual shells or their halves.The mean shell weight changes exponentially with shell length, from 6.3 mg at the smallest class (4 to 5.9 mm) to 1182.1 mg in the largest class (34 to 35.9 mm).The 400 mg sample weight used in the assays makes it possible to analyse single halves measuring 24 mm and more.
The derivatograms obtained make it possible to calculate the amounts of physically bound water, organic matter, and calcium carbonate in the Dreissena shells.All the derivatograms obtained do not differ significantly from one another, therefore a single derivatogram is presented only (Fig. 2).The most legible curve is the DTG one; its interpretation will serve as a basis for further considerations.The DTA curve which incorporates both endo-and exothennic effects is less legible, particularly the part that records the transfer between the loss of physically bound water and organic matter combustion.The derivatogram is presented in its original, unchanged form, hence different lengths of various sections in spite of the assumed constant temperature increase.The constant temperature increase is seldom attained in practice and depends on both the properties of a substance tested and characteristics ofthe apparatus used.
The DTG curve shows three effects of mass loss.The first is related to the loss of the physically bound water and occurs at 25 -30°C to 130 -170°C with a maximum at 70 -90°C.This effect is interpreted as a result of dehydratation.The water content is equal to the weight loss as read from the TG curve.The second, broad effect is related to organic matter combustion and proceeds at 130 -170°C to 539 -580°C with a maximum at 283 -302°C.The weight loss related to this effect and read from the TG curve identifies the organic matter content.More precisely, the content of solids remaining after the organic matter combustion is regarded as negligible and confined to the measurement error limits.The third, very conspicuous effect is brought about by CaC0 3 breakdown as in the reaction:  Thee:ffectisvisibleat535 -580°C to 920-930°C with a maximum at 895-905°C.The weight loss calculated from the TG curve results from C0 2 release and is multiplied by the gactor of2.27 4 ( CaC0 3 : C0 2 = 2.27 4) to obtain the CaC0 3 content.
The dehydratation reaction is reflected in the ])TA curve as a poorly visible endothermic effect smoothly passing into organic matter combustion.The latter shows two peaks: the first at 328-355°C and the other at 435-465°C.While the latter is more pronounced on most derivatograms, the first is related to the principal weight loss effect (DTA curve).The CaC0 3 breakdown corresponds to anextensiveasymmetriceffectontheDTAcurvewithamaximumat900-910°C.
Results of the derivatographic analyses are presented in Table 1.The content of physically bound water was found to range within 0.4-0.8% with a mean (15 samples) of 0.5%.The presence of water in the samples, in spite of drying them at 105°C, can be explained by the water absorption properties of the substance tested, the samples absorbing moisture while being prepared for the derivatographic analyses.The water content is disregarded in the further discussion.
The minimum and maximum CaC0 3 contents were 88.7% (4.0-7.9 mm class) and 96.4% (26.0 -27.9 mm and 28.0 -29.9 mm classes), respectively, the mean content amounting to 93.4%.The shell length versus CaC0 3 content correlation was positive (r = 0.867).The regression equation showed a shell length increase by one milimetre to be accompanied by a CaC0 3 content increase by 0.21 %.
The combined contents of organic matter and CaC0 3 ranged from 92.3% ( 4.0 -7.9 mm class) to 98.4% (26.0 -27.9 mm and 28.0 -29.9 mm classes) with a mean of 95.9%.The correlation between the two components was clearlynegative(r=-0.865).Organic matter and CaC0 3 are the basic components of each shell; an increase in the content of one of them brings about a decrease in the other.The regression equation showed a 1% increase in the CaC0 3 content to be accompanied by a decrease in organic matter by 0.21% (Fig. 4).
The organic carbon (Corg) content of Dreissena shells was found to vary from 0.4% (26.0-27.9mm length class) to 1.22% (8.0-9.9 mm class) with a mean (six samples) of 0.61% (Table 2).Due to a small sample size, the Corg versus shell length relationships were not calculated.On the other hand, Corg correlated positively with organic matter (r = 0.993).The corresponding regression equation showed a 1% increase in the organic matter content to be accompanied by the organic carbon content increasing by 0.48% (Fig. 5).The organic matter to corg ratio is not constant and varies from 2.95 to 5.00.Based on the regression equation for the two components, the ratio may be given as: organic matter A relationship between Corg and organic matter residue after the removal of Corg (organic matter-Corg) was calculated as well.The correlation coefficient was r = 0.983.The corresponding regression equation showed a 1% increase in the content of Corg contained in organic matter to result in an increase of the remaining organic matter components by 1.04% (Fig. 6).
Table 3 summarizes data on contents of 13 chemical components of Dreissena shells.As a single analysis only was made (a 24.0-25.9mm length class sample), the data should be treated as preliminary.The Si0 2 content was 1.4% and is interpreted as a contamination by mineral components of bottom sediments.The contents ofK and Cd are given as zero values.As the elements were determined to 0.001%, the zero values should be taken as a possibility ofK and Cd occurring in shells in amounts lower than 0.001%.Contents of the remaining elements varied from 0.001% (Ni) to 0.203% (Fe).The combined content of these components (without Si0 2 ) was 0.452%.
It is interesting to compare elemental composition ofthe Dreissena shells with that of unconsolidated sediments Dreissena populations dwell on.Mean values of three sediment samples (0 « 0.25 mm fraction) are presented.The concentration coefficient (KA) was calculated as a ratio between the mean shell content and the mean sediment content of an element.As Si0 2 in the Dreissena shells was assumed to be a contamination the concentration coefficient was not calculated in this case.Values of the coefficient for the remaining elements were very low (« 1) except for Cu (KA = 1).
Another coefficient of concentration, KB, was calculated as a ratio between the mean shell content of an element and its clarke.The KB values for Pb, Co, Zn, and Cu were found to range from 2.00 to 4.33, the values being less than 1 for remaining elements.

Discussion
In view of the techniques used to prepair the materials for assays, the results presented in this paper are treated as mean values for the Dreissena shell length classes separated.
Statistical relationships presented are regarded as tentative only, to be verified by analysing a larger body of data.The conclusions resulting from interpretation of the regression equations should be treated as expected values which, in certain cases, may deviate from the true ones.Several derivatograms were obtained for a single homogenous sample to verifY the interpretation of the results of the derivatographic analysis.The between-measurement differences were within the experimental error limits and amounted to 0.1 %.The criteria of the derivatogram interpretation are given clearly in the literature.For the purpose of this study the guidelines established by WYRWICKI (1988) are followed.However, it remains to be decided whether the organic matter content, determined in this way, may be equalled with the shell conchiolin content.WYRWICKI (1988) questions the validity of comparing the two substances.A bivalve shell is composed of three layers.The outer layer (periostracum) is made of organic matter called the conchiolin.It seems that the problem of comparing the devaritographically determined organic matter with conchiolin can be solved by employing other techniques, mainly X-raying.
The organic matter versus Corg content relationship presents an interesting problem.It seems probable that a low number of data has in some way affected the relations between the two components as presented in this paper.Nevertheless, the contents of the two components did not increase in proportion to each other.With the increase in both organic matter and Corg contents, their ratio was observed to decrease.Both substances were, without doubt, determined correctly.Generally, organic matter consists of organic carbon bound to elements such as hydrogen, nitrogen, and oxygen.As no proportion between increasing amounts of organic matter and Corg could be observed, a change in organic matter chemistry in terms of altered numerical relations between the components is a likely explanation.It is consequently assumed that Dreissena shells contain compounds other than corg which produce thermal effects similar to those of c org and are responsible for the disproportionate increase of the two components discussed.
In order to grow, a Dreissena shell needs certain chemicals.The chemistry of the shell reflects, to some degree, that of its environment.Unconsolidated lacustrine sediments form an active system in which proceed a number of biologically important processes.These processes provide for, and control transformation of numerous biogenic elements.Thus the coefficient of concentration KA is interpreted as a measure of an element's ability to migrate out of the sediment into the shell structure.The very low values of the coefficient found may evidence an inhibited sediment-shell migration of all the elements except copper in which case the value of the coefficient may suggest an enhanced incorporation into the Dreissena shell.
The other coefficient calculated, KB, compares the elemental content ofthe shell with that of the lithosphere (clarkes).Thus clarkes of various elements can be regarded as expected values typical of the environment.Four elements: Pb, Co, Zn, and Cu are concentrated in the Dreissena shells in amounts exceeding their clarkes.The contents of the remaining elements are lower than their clarkes.It was not possible to calculate the KB value of Cd as the assays were made to lxl0-3 %, while the Cd clarke is 5x10-5 %.
The contents of several elements in the Dreissena shells are relatively low.
However, if their total amounts deposited in the shells of the entire Dreissena population in a water body were calculated, the values arrived might be quite substantial.
The chemistry of D. polymorpha shells was earlier presented by WIKTOR (1969).A direct comparison with his results is difficult because WIKTOR (1969) analysed pulverised unsorted bivalves 3 -30 mm long.the results thus obtained describe the averaged chemical composition of a whole Dreissena individual (shell + soft parts).The shell chemistry versus length relationship presented in this paper may cast doubts on the validity of calculating amounts of CaC0 3 and other compounds, accumulated in the Dreissena shells, by the use of those averaged values.

Conclusions
The derivatographic analysis is a relatively simple and rapid method of simultaneous quantifying of several substances, provided these are thermally active and occur in appropriate amounts in the samples.The "appropriate amount", understood as a minimal quantifiable amount, as different for different substances and depends both on thermal properties of a substance and on thermal properties of the remaining compounds making up the sample and and their proportions in it.For example, assuming that the whole Mg amount determined in the Dreissena shell (Table 3) is bound as MgC0 3 , the calculations yield 0.19% MgC0 3 .This is too low for the derivatogram to record the thermal breakdown.
However, the method can be used to determine, with reasonable accuracy comparable to that of chemical procedures, contents of the basic chemical components, viz.organic matter and CaC0 3 .For the purpose ofthis study, 400 mg samples were analysed.Much smaller samples, e.g.40 mg, can be used as well, in which case referring to Fig. 1 -single shells or their parts can be analysed.Further work employing the derivatographic analysis will hopefully make it possible to select optimal procedures for studying molluscan shells.
The results of assays of the remaining chemical components, obtained with other analytical techniques, provide insights into the chemical composition of Dreissena shells and enable calculations of coefficients of concentration for various elements.
Further studies on molluscan shell chemistry, employing new research techniques and carried out in chemically different ecosystems may contribute to elucidation ofthe course of geochemical reactions in the respective environments, with a particular emphasis on relations between chemistry of shells and that of bottom sediments and water.
The length classes up to 20 mm contained 100 shell halves each,

Fig. 1 .
Fig. 1.Dreissena shell half mean weight distribution in length classes

Fig. 3 .Fig. 5 .
Fig. 3. Distribution of contents of CaC0 3 , organic matter and Corg in Dreissena shells and regression equations for contents versus shell length relationships

Table 1 .
Chemical composition (weight %) of Dreissena shells as shown by derivatographic Organic matter and Corg content (weight%) and organic matter: Corg ratio inDreissena