Establishment of a sampling strategy for the use of blue mussels as an indicator of organotin contamination in the coastal environment Koji Shindo and Akira Otsuki* Department of Marine Science and Technology, Tokyo University of Fisheries, 4-5-7 Konan, Minato-Ku, Tokyo 108, Japan. E-mail: md97012@edufs.tokyo-u-fish.ac.jp (Koji Shindo); akirao@s4201.tokyo-u-fish.ac.jp (Akira Otsuki); Fax:+81 3 5463 0398; Tel:+81 3 5463 0458 Received 10th December 1998, Accepted 16th March 1999 The application of the Mussel Watch concept to the assessment of chemical contamination in the coastal environment is still premature, since the relationship between the physiological and ecological aspects of blue mussels and the accumulation of contaminants in their soft tissues remains unclear. We cannot yet directly estimate with known confidence the degree of chemical contamination from the levels of contaminants in the soft tissues of mussels. An understanding of the source and range of variability in the tissue concentrations of contaminants is essential, and the establishment of a biomonitoring sampling strategy to minimize the eVect of identified sources of variability is required.The present study was conducted to clarify the characteristics of organotin accumulation in blue mussels under various conditions in Tokyo Bay, and to establish an optimized sampling strategy of mussels as exposure indicators of organotin contamination in Japan. It was clear that the sample number, individual size, spawning activity and vertical habitat were factors causing a variation of tissue concentration.Based on a quantitative estimation of the variability of organotin concentration in mussel tissues under various physiological and natural conditions, we suggest that a composite sample of 30 mussels (3–5 cm in shell length), collected from the infralittoral zone prior to their spawning season, is essential to reduce the variability between individuals and to obtain reproducible analytical values at a sampling site.For monitoring sites where natural blue mussels cannot be collected, an alternative method of transplanting blue mussels from a relatively clean area would be required. accumulation in mussel tissue. Little research on the physio- Introduction logical and ecological factors aVecting the bioaccumulation of The Mussel Watch concept, based on the bioaccumulation of contaminants in blue mussels has been performed.As a contaminants by natural blue mussels, is becoming a worldwide consequence, we cannot yet completely validate the origin and monitoring system of chemical contamination in the coastal range of variability of contaminant concentration in biological environment, because the analysis of contaminants in mussel tissues, or the spatial or temporal coverage of contamination tissues is regarded as a superior way of providing information expressed by only one value of tissue concentration.The about the occurrence and distribution of contaminants.1,2 answer to the origin and range of variability of tissue concen- Nowadays, many studies are providing us with a rough tration is directly connected to the establishment of a sampling approximation of the status and trends of contaminants in the strategy that will minimize the eVect of identified sources of coastal environment.3–6 The genus Mytilus consists of a group variability. An optimized sampling strategy should provide us of widespread sedentary mussels.Many contaminants accumu- with a method of collecting a ‘reliable sample’ that reflects the late in their soft tissues to high concentrations, making the contaminant levels in ambient seawater for the assessment of chemical contamination.Thus, such a strategy is also essential measurement of such chemicals easier than in seawater for clarifying the limits of its applicability as a biomonitor- samples.The tissue concentration of contaminants can be ing method. regarded as a time-averaged or time-integrated concentration The purposes of this study were to clarify the characteristics for the ambient environment. This implies that the Mussel of organotin accumulation by blue mussels under various Watch programme may become a useful coastal monitoring conditions in Tokyo Bay, Japan, and to establish an optimized method.However, the application of biomonitoring using blue sampling strategy of mussels as an indicator of exposure to mussels to assess chemical contamination has several important organotin compounds in Japan. The optimized sampling strat- problems in spite of careful sampling and chemical analysis: egy and its applicability as a monitoring method for six (i) taxonomic diYculties in sampling; (ii) lack of prescribed organotins were considered in terms of the usefulness of this pertinent sampling/analytical strategies; (iii) uncertainties in technique as an analytical method for the determination of variability and spatial or temporal resolution of tissue concenorganotin compounds in mussel tissue and with regard to the tration; and (iv) limitations of the approach itself.7,8 The most variability of tissue concentration due to the ecological aspects popular strategies of the National Oceanic and Atmospheric of blue mussels.Administration (NOAA) in the USA and the Re�seau National d’Observation (RNO) in France are no exception to these Materials and methods problems. The most basic issue in these problems seems to be a lack of understanding of how important these factors are in Sample collection and preparation the production of reliable data, although it is well known that a variety of factors influence the relationship between the Natural blue mussels were collected by hand from three stations in Tokyo Bay, located in the central part of Japan concentration of contaminants in ambient water and the J.Environ. Monit., 1999, 1, 243–250 243(Fig. 1): Tokyo-Light, located oVshore, is an artificial construction supported on a steel pole; Odaiba is an artificial rocky reef in the recesses of the Bay; Kannonsaki is a natural rocky reef located in the mouth of the Bay. Blue mussels were sampled mainly in June, August, November, February and April from 1994 to 1996, based on their ecological characteristics in Tokyo Bay (Fig. 3, see later). The identification of blue mussels was based on the characteristics of the shell shape and colour of the mantle edge and shell.9 Mussels were collected as a composite sample of individuals from several single clumps within a few square feet at the lower tidal level (±0 m). We selected 20–150 mussels per analytical sample according to the objectives.The samples were kept in cold storage (<10 °C) until pretreatment in the laboratory. Pretreatment was carried out within the day of sampling. Water retained inside the shell was completely removed with a stainless strainer, the byssus was discarded and the mean whole wet weight of the sampled mussels was measured.Mussels from each sampling site and date were divided into six size groups according to shell length: 1.5–1.8 cm, 1.8–3 cm, 3–4 cm, 4–5 cm, 5–6 cm and 6–7 cm. Fig. 2 The equipment for transplanted bivalves. After taking out the soft tissue, the adhering water was fully removed by draining for 20 min in a stainless strainer. For the study on organotin distribution in the soft tissues, 50–70 individuals (3–5 cm in shell length) were dissected into gill, ation and gas chromatography with flame photometric viscera, mantle and muscle.Finally, each sample was separately detection. Details of the analytical procedure are given in homogenized (10 000 rpm, 10 min) and then frozen (-40 °C) previous papers.12–16 until analysis. Details of sample handling, pretreatment and Each organotin peak in the environmental samples was storage followed the previous papers.10,11 verified by the addition of organotin standards.Based on the internal standard method, the concentrations of the six ana- Transplantation procedure lytes in the extract were calculated by comparing the peak area of the spiked internal standard [tripentyltin chloride Natural blue mussels for transplant were collecteby hand (TPeT)] with that of each analyte.Organotin values in the from a relatively uncontaminated site in Tokyo Bay (Iwai or extract were expressed as the chlorides. The limit of detection Tateyama) or a seashore rocky reef facing the Pacific Ocean (S/N=2) was 0.5–0.7 ng (chloride) g-1 wet weight for the six (Kujukuri) outside Tokyo Bay.The mussels were kept in organotins in mussel samples. The reported concentration was uncontaminated seawater for about 1 month before transplanusually expressed as a mean value of three determinations tation. They were transplanted to Tokyo-Light and Hukuura from one extract, and was corrected by the reagent blank and Harbour. We transplanted 30 individuals (3–5 cm in shell recovery of each organotin species.length) into each of five positions with respect to tidal height for about 1 month in autumn and winter from 1994 to 1997 (Fig. 2). Pretreatment of tissue samples was performed in the Results and discussion same way as for natural mussels. Evaluation of the analytical method for the determination of organotin compounds in blue mussel soft tissue Analytical procedure Recovery was validated by comparing the results of six repli- The determination of six organotin compounds [tributyltin cate analyses of an uncontaminated sample spiked with (TBT) and triphenyltin (TPT) and their degradation products] 16–28 ng (chloride) g-1 wet weight (this value was the average in homogenized soft tissue (3 g) of blue mussels was carried level of samples from Tokyo Bay) with those obtained from out using sonication with tropolone–benzene, propyl derivatizthe analysis of propylated standards.Table 1 indicates the recovery rate of 10 organotin chlorides in blue mussel tissue through the whole analytical procedure. The recovery by our procedure was in the range 56–116% for butyltin and phenyltin compounds. Repeatability was assessed with five replicate analyses of three average contaminated samples.The relative standard deviation (RSD) ranged from 9.4 to 26% for the 10 analytes and was 8.2% for TPeT. The accuracy of our analytical procedure was validated by five replicate analyses of 2 g of certified reference material from the National Institute for Environmental Studies (NIES No. 11 Fish Tissue).Our method yielded 69.9±7.2% of the TBT certified value and 50.4±8.6% of the TPT reference value. It should be noted that the physicochemical form of the analyte in the spiked sample is probably diVerent from that in the real sample, especially in the analysis of biological tissues where the organotin compounds may be bonded to lipids or proteins, and that the use of spiked samples can lead to an overestimation of the extraction eYciency.17 Thus, the Fig. 1 Sampling stations in Tokyo Bay. values of the recovery for spiked samples may be higher than 244 J. Environ. Monit., 1999, 1, 243–250Table 1 The recovery of organotin compounds in bivalve samples by our analytical procedure MMT DMT TMT MBT DBT TBT MPT DPT TPT Recovery (%) 135.7 80.4 46.2 115.9 100.5 81.0 56.0 111.8 95.5 RSD (%) 13.1 13.8 25.6 20.7 18.9 14.8 26.1 9.9 13.2 Number of analyses n=6.TPeT: absolute recovery, 79.0%; RSD, 18.2%. MMT, monomethyltin; DMT, dimethyltin; TMT, trimethyltin; MBT, monobutyltin; DBT, dibutyltin; TBT, tributyltin; MPT, monophenyltin; DPT, diphenyltin; TPT, triphenyltin. those in real soft tissues, and our analytical results might be temporal bias [Table 2(a)].In spring, blue mussels less than 1.8 cm in shell length were few and, in summer, many mussels 30–50% lower than the concentrations in real samples. Judging from the recovery of spiked samples using a highly longer than 5 cm disappeared. This seasonal population change suggests that mussels less than 1.8 cm in shell length reproducible analytical method, our analytical procedure is a reliable method for organotin analysis in mussel soft tissues, and those longer than 5 cm may not always be sampled, while blue mussels of 3–5 cm in shell length can be collected although the results obtained from the recovery test may not completely reflect the amount of organotin compounds throughout the year.Thus, the number and size of accessible mussels on each sampling date may be greatly limited by extracted from real tissue samples.The analytical variability was evaluated to be <5–26% RSD of the tissue concentration. their population fluctuation at each sampling site. Table 2(b) shows the seasonal soft tissue weight of individuals in our sampling stations; the seasonal variation was within the range Sampling of natural blue mussels 2–20%.Our results based on the population dynamics and Although mussels are dominant organisms in the intertidal average weights of individuals of naturally occurring blue sessile fauna in Tokyo Bay, taxonomic identification of Mytilus mussels suggested that individuals of 3–5 cm in shell length during sample handling enabled us to discriminate easily between blue mussels and other mussels and bivalve genera.Blue mussel samples could easily be collected from each sampling site. However, we cannot yet discriminate between Table 2 Population dynamics and sampling availability of natural two blue mussel species (M. galloprovincialis and M. edulis), mussels at sampling stations using morphological characters, allozyme characters and reproductive patterns.Blue mussels in Tokyo Bay were tenta- (a) tively identified as Mytilus edulis galloprovincialis.18 Fig. 3 Sampling availability of natural mussels shows the relationship between the ecological characteristics and sampling time of blue mussels in Tokyo Bay. Their Individual size/cm ecological characteristics indicate that their physiological condition and population size change drastically within a year, as Sampling date 1.0–1.5 1.5–1.8 1.8–3 3–4 4–5 5–6 6–7 7–8 a result of the intense heat on summer days, spawning activity June ’94 + + ++ ++ ++ + - N and recruitment of young mussels.18–20 Aug.’94 + + + +++ N N N The spatial and temporal distribution, individual size and Nov. ’94 - ++ ++ + + N N N volume of soft tissue of blue mussels under various natural Jan.’95 - ++ ++ + + - N N conditions were examined as primary factors in sample Feb. ’95 - + ++ ++ + - N N handling, because the Mussel Watch concept depends on the Apr. ’95 ++ + ++ ++ ++ - - N presence of natural mussels. Our observation of the shore June ’95 + ++ ++ +++ + -N Aug. ’95 + ++ ++ +++ - -N surrounding Tokyo Bay indicated that the distribution of Jan. ’96 - - + + - - N N blue mussels was patchy, although their population was large Feb.’96 N N + + - - N N in each habitat. This means that we cannot necessarily collect May ’96 + + + ++++- -N blue mussel samples in locations where we need to monitor N, none present or no sampling; -, few present (need for hard contaminants. The seasonal size composition of blue mussels sampling); +, enough present for sampling; ++, many present.in the lower tidal level of our sampling stations indicated that the presence of small and large individuals had a (b) Ratio of average weight to the value of 4–5 cm natural mussel samples Individual size/cm Sampling date 1.5–1.8 1.8–3 3–4 4–5 5–6 6–7 June ’94 0.07 0.22 0.50 1 (9.5)a 1.6 2.3 Aug. ’94 0.10 0.22 0.76 1 (6.9)a N N Nov. ’94 0.07 0.21 0.56 1 (9.3)a N N Jan.’95 0.09 0.13 0.62 1 (9.1)a 1.4 N Feb. ’95 0.07 0.19 0.61 1 (9.9)a 1.3 N Apr. ’95 0.03 0.23 0.64 1 (8.8)a 1.5 2.4 June ’95 0.05 0.20 0.54 1 (9.1)a 1.6 2.3 Aug. ’95 0.07 0.18 0.60 1 (8.2)a 1.2 1.9 Jan. ’96 0.05 0.11 0.42 1 (3.6)a 1.8 N Feb. ’96 N 0.18 0.50 1 (2.8)a 2.0 N May ’96 0.02 0.18 0.55 1 (3.5)a 1.6 3.2 aThe mean of actual wet weight (g wet). N, none present or Fig. 3 Ecological characteristics and sampling time of natural blue no sampling. mussels in Tokyo Bay. J. Environ. Monit., 1999, 1, 243–250 245were the most average-sized adults growing steadily in rapidly and intestinal tract from ambient water and food. A strong aYnity of organotin compounds with mucin in gills has also changing populations. Judging from the average wet weight of individuals, the been reported.22 It is suggested that organotin compounds may primarily accumulate in gills and viscera, and be second- minimum number of individuals required for repeatable analysis of the six organotins was a composite sample of more than arily transferred into lipid- and protein-rich tissues, mantle (including gonadal tissue) and muscle by circulatory transport.five mussels of 5–6 cm in shell length, more than 12 mussels of 3–4 cm in length or more than 150 mussels smaller than The four diVerent tissues showed almost the same accumulation patterns of TBT, TPT and their degradation products 1.8 cm in length (Table 3). However, BeliaeV et al.3 and Daskalakis et al.21 stated that a suYcient number of individuals in all seasons.This similarity suggests that blue mussels do not metabolize the tri-form of organotin compounds to di- to average the residual diVerences between individuals was a composite sample of more than 30 bivalves, and that the use and mono-forms in soft tissues. This suggestion is in agreement with the finding that the detoxifying activity of mussel tissue of more than 30 individuals from each population is suYcient to permit eVective comparison (detect diVerences of 40%) of for organotin compounds is very low.23,24 A significant decrease in the level of accumulation of the tissue concentrations in these two instances.Therefore, on the assumption that more than 30 individuals of the same shell six organotins in mantle tissue during the spawning period was observed, although the accumulation pattern and level of length collected from a single clump within a few square feet in the infralittoral zone is enough to average the residual the six compounds in the four diVerent tissues did not indicate a definite seasonal variation.The decrease in accumulation diVerences between individuals, and that all collected individuals been exposed to a similar environmental and physiological level showed the same tendency as the change in the gonad index at spawning time.Previous reports25–27 demonstrated conditions, we focused our discussion on two factors: (i) the characteristics of organotin accumulation in blue mussels; and that many contaminants are incorporated into mussel gonad or bonded to lipids and proteins in gametes, and that fluctu- (ii) the establishment of an optimized sampling strategy.Moreover, we individually estimated the variability associated ations in the accumulation of contaminants were correlated with the seasonal cycle of gametogenesis and spawning activity. with sampling date, sex diVerence, individual size and habitat height with application of the optimized sampling strategy for This implies that organotin compounds accumulated in the gonad (or mantle) are released with eggs and sperm during three parameters except the investigated parameter.The weight of variability due to each investigated parameter was statisti- spawning (Fig. 4). Organotin compounds released with gametes were estimated at 12–63% of their accumulated con- cally calculated by relative comparison between the analysed values of the six organotins in an optimized mussel sample centration in mantle tissue.Errors due to sampling date, especially in the spawning period, were estimated to be and in mussels under the influence of the investigated ecological parameter. 2.5–18.4% of the analysed values derived from a reliable mussel sample prior to the spawning season. Accumulated concentration in each organ under various physiological conditions DiVerences in accumulated organotin concentration due to sex, shell length and vertical habitat The influence of the reproductive and other seasonal cycles on the accumulated concentration in each organ of the mussels Sexual diVerence in accumulated concentration.On each sampling date, we collected individuals whose sex was discern- was examined by determining the accumulated organotin concentration in 50–70 individuals of 3–5 cm in shell length.ible by the colour of the mantle. We determined the accumulated organotin concentration in 50–70 (male or female) First, the physical condition and reproductive activity of the mussels were observed under natural conditions.In summer individuals of 3–5 cm in shell length. Mantles of female and male mussels showed almost the same accumulation pattern and just after the spawning season, soft tissues obviously lost total weight [3–14%; Table 2(b)]. The composition of the of TBT, TPT and their degradation products under seasonal and reproductive cycles (Fig. 4). Variability due to sex diVer- mantle including gonadal tissue showed a seasonal variation (minimum, 9%; maximum, 30%).The gonadal index calculated ence was estimated to be 1.4–7.8%. With regard to the accumulation of organotin compounds in soft tissue in blue in blue mussels of 3–5 cm in shell length showed the same tendency as that observed previously.18 Almost all mature mussels, it is clear that sex is not an important variation factor.individuals seemed to have finished releasing gametes by April. However, we could diVerentiate the sex of some individuals in Organotin accumulation levels in mussels of diVerent size. Seasonal accumulation levels for diVerent shell lengths under the population by the colour of the mantle not only from mid- October to March, but also in summer. various physiological conditions were examined by determining the accumulated organotin concentrations in 25–150 individ- The accumulated concentration of organotin compounds in each tissue increased in the order of gill, viscera, mantle and uals of each shell length.All mussels had higher levels of organotin compounds in their tissue than those present in muscle; gill and viscera concentrations were 1.2–4 times higher than those of other tissues, showing that the mussels accumu- ambient seawater.However, no relationship was found between accumulation level in whole soft tissues and shell late high concentrations of the six organotins in gill, viscera Table 3 Sample collection of natural blue mussels Shell length/cm 1.5–1.8 1.8–3 3–4 4–5 5–6 6–7 3–5a Average weight of 0.48–0.72 1.0–2.1 2.7–6.4 6.1–9.2 9.4–16 15–22 4.6–9.2 individuals/g wet Minimum number for 150 30 12 10 5 3 70 repeatable analysis Number of collected mussels 150 100 50 30 25 25 100 aSample of various organs. 246 J. Environ. Monit., 1999, 1, 243–250Fig. 4 Organotin concentration in mantle including gonadal cell and gonad index of natural mussels. length (Table 4). Accumulation levels in mussels of diVerent Accumulated organotin concentration in diVerent vertical habitats.The diVerence in vertical position of mussel habitats shell lengths did not show significant seasonal variation. Small mussels, 1.5–1.8 cm in shell length, often had 1.1–3 times relates to the immersion time in seawater. Judging from the average condition of Tokyo Bay and the tidal level data higher levels of organotin compounds than those in mussels of 3 cm or above.Mussels of 3–5 cm in shell length (1–2 years calculated by the harmonic analysis of tides, the submerged ratio of mussels in the infralittoral zone is 0.5–10% of sub- old) showed nearly the same concentration level and accumulation patterns of organotin compounds in whole soft tissue merged time in the intidal zone (Fig. 6). The drying of mussels when out of seawater negatively aVects their growth rate and on every sampling date. Compared with large adult mussels longer than 3 cm in shell reproductive activity. The growth of mussels in our samples before and after transplantation showed the same tendency as length, mussels of less than 1.8 cm being recruited to mussel beds in each spring show rapid feeding and a high growth in the previous reports.18,19 Compared with adult (3–5 cm) mussels before transplantation, those transplanted to the tidal rate, but a low excretion rate of extraneous substances.19,28,29 An absence of correlation between shell length and lipid zone (above the mid-tide level ) showed some mortality and 60–110% growth in wet weight of the survivors; those trans- content has also been reported.30 The biological characteristics of small mussels ( less than 1.8 cm long and less than 1 year planted to the infralittoral zone (beneath the mid-tide level ) showed 70–140% growth in wet weight (Fig. 6). Transplanted old) would lead to a high accumulation of contaminants. Moreover, the eVect of reproductive activity on the accumu- mussels showed little increase in shell length after transplantation.lation of organotin compounds in soft tissue is significant before and after spawning time, because there is a large We determined the accumulated organotin concentrations in 30 individuals of 3–5 cm in shell length transplanted to five diVerence in accumulation characteristics between mature adult ( longer than 3 cm) and immature ( less than 1.8 cm long) habitats at diVerent tidal heights.The results showed that the habitat position with respect to the tide caused significant mussels. The variability due to individual size was estimated to be ±6–38% for TBT and 10–116% for TPT of the analysed diVerences in accumulated organotin concentrations. Mussels in the infralittoral zone (from average lower tidal level to value derived from 3–5 cm mussels.The population dynamics and average weight of individuals [Tables 2(a) and 2(b)] sublittoral zone) were 1.2–1.5 times stronger accumulators than those in the intidal zone (Table 5). However, the mussels suggested that the 3–5 cm mussels were the most ‘average’ adults. Mallet and Carver31 and Lowe et al.32 reported that in these five diVerent tidal levels showed nearly the same accumulation patterns of the six organotin compounds in the normal 1–2-year-old mussels have a stable internal body condition for growth and reproduction under rapidly changing whole soft tissue under various seasonal conditions.The eVect of an increase in shell length on the accumulation level can be natural conditions, and indicated the similarity in reproduction and internal body condition between such blue mussels.Our disregarded, because our long-term field study made it clear that the uptake of organotins in the soft tissue of transplanted results suggest that the similarity of level and pattern of organotin accumulation in mussels of 3–5 cm in shell length blue mussels was rapid and that the apparent equilibrium of organotin concentration between the soft tissue and ambient in our sampling sites is dependent on the similarity in reproduction activity and internal body condition of the mussels seawater required only 2–3 weeks.33 The diVerence in height of the habitats resulted in diVerent immersion (or exposure) (Fig. 5). Judging from sample handling and the similarity of accumulation in tissue, more than 30 mussels of 3–5 cm in times in contaminated seawater.The variability due to habitat height was estimated to be ±10–27% for TBT and 2–64% for shell length are suitable as a composite sample for use as an exposure indicator. TPT of the average value for mussels from the lower tidal level. This variability suggested that the immersion time in seawater aVects not only the advantage of habitat but also the Table 4 Variability of organotin concentration in diVerent sized accumulation level of organotins.natural mussels in Tokyo Bay Optimized sampling strategy Ratio of concentration to the value of 3–5 cm sample Currently, the most popular programmes, those of the NOAA34 and RNO,6 have a regularized method to collect 1.5–1.8 cm 1.8–3 cm 3–5 cm 5–6 cm 6–7 cm ‘reliable samples’. NOAA’s sampling strategy attempts to limit environmental influences by collecting annually a set number TBT 1.38 1.23 1 (19.3)a 0.84 0.89 (30 individuals ×3) of blue mussels with a set size range TPT 1.13 1.24 1 (4.48)a 0.90 2.16 (5–8 cm) at the same time (in winter prior to the spawning aThe mean of actual concentration value (ng g-1).season), at the same depth (continuously submerged mussels) J.Environ. Monit., 1999, 1, 243–250 247Fig. 5 Seasonal variation of organotin accumulation in mussels of 3–5 cm in shell length. Fig. 6 Actual submerged ratio of mussels in five vertical habitat positions and the growth of mussel samples before and after transplantation. Table 5 Variability of organotin concentration in mussels transplanted to five diVerent tidal heights in Tokyo Bay Ratio of concentration to the value of sample in average tidal level Upper limit of Middle of Average of lower Sublittoral Lower limit of distribution tidal level tidal level zone distribution TBT 0.896 0.732 1 (72.4)a 0.939 0.827 TPT 1.02 1.45 1 (6.58)a 1.64 1.21 aThe mean of actual concentration value (ng g-1).and in the same site every year.The purpose of the composite collect native blue mussel samples in locations where measurements are required to monitor contaminants; (ii) population samples in one sampling site is to minimize sample variability. To reduce variability due to spawning state and hydrography, fluctuations within a sampling site greatly limit the number and size of accessible mussels on each sampling date; (iii) the all sampling has to be made within±3 weeks of a certain date every year.RNO’s sampling strategy attempts to take within- reproductive cycle aVects the accumulated concentration, especially due to the significant decrease in accumulated year fluctuations in chemical concentrations into account by collecting seasonally a set number (50 individuals, a number organotins in mantle tissue during the spawning season, while blue mussels show nearly the same accumulation pattern in high enough to minimize the eVect of individual variations) of blue mussels with a set size range (3.5–6 cm) at the same site all seasons; and (iv) diVerences in immersion time in seawater resulting from habitat height generate diVerences in exposure (intertidal zone) in autumn every year.However, these regulations only indicate sample number, size, sampling location, time to organotins. Based on the eVect of these ecological aspects when mussels are used for monitoring organotin date of sampling and the reasons to regularize sampling. Therefore, we clarified the characteristics of organotin accumu- contamination, we propose a sampling strategy for blue mussels as exposure indicators in Japan as follows.lation in terms of the ecological aspects of mussels under various natural conditions. (i) Sampling location: a large number of individuals must be collected from several clumps within a few square feet at Our quantitative estimates of the variability of organotin concentrations in diVerent mussel samples under natural con- the same depth (continuously submerged mussels), such that all mussels essentially receive similar environmental stress and ditions suggest the eVect of ecological aspects when blue mussels are used as exposure indicators: (i) we cannot always exposure conditions. 248 J. Environ. Monit., 1999, 1, 243–250(ii) Sampling date: to reduce variability due to spawning, mussels not only restricts the number and size of accessible samples, but also prevents sampling in certain locations where samples must be collected from the same sampling locations at the same time, prior to the spawning season. contaminants should be monitored; (ii) we cannot yet directly estimate with known confidence the degree of chemical con- (iii) Sample number: a composite sample of more than 30 individuals from each population is required for satisfactory tamination based on the levels of accumulation of contaminants in the soft tissues of mussels.Further studies based on reduction of sampling variance. (iv) Size of mussels: collected mussels should have the an optimized sampling strategy should reveal with known confidence the temporal and spatial coverage of ambient narrowest possible range of shell length in order to reduce satisfactorily ecological and physiological diVerences between contamination drawn from mussel tissue concentrations, and the capabilities and limitations of the Mussel Watch concept individuals (an average length of 3–4 cm or 4–5 cm presents least variability).as a monitoring method for chemical contamination.(v) Availability of samples: some requests for contaminant monitoring may be in locations where no natural mussels are Acknowledgements living; we must standardize a transplantation method as an alternative for such cases. We thank the staV of Tokyo-Light, Banda Field Marine Advances in analytical instrumentation enable us to analyse Laboratory and the National Institute of Fisheries Science for trace levels of many contaminants in biological tissues.We their support. We also give special thanks to Drs. Hiroshi can solve analytical problems by the spread of standardized Yamakawa and Osamu Oku for their advice. analytical procedures and certified reference mussel tissue material. Thus, it is increasingly recognized that the major risks of error in environmental monitoring are not in the References laboratory (analysis of contaminants) but rather during field 1 M.E. Gurtz, in Biological Monitoring of Aquatic Systems, ed. S. L. operations (sampling) or prior to analysis (sampling, storage Loeb and A. Spacie, Lewis, Florida, 1995, pp. 345–347. and sample pretreatment).10 Pretreatment of mussel samples 2 E.Goldberg, Environ. Monit. Assess., 1986, 7, 91. after collection produced another suggestion for the optimized 3 B. BeliaeV, T. P. O’Connor, D. K. Daskalakis and P. J. Smith, sampling strategy. The analysis of only one organ of blue Environ. Sci. Technol., 1997, 31, 1411. 4 G. G. Lauenstein, Mar. Pollut. Bull., 1995, 30, 826. mussels as an exposure indicator should be avoided, because 5 M.Morita and Y. Shibata, Japanese–French Workshop on Recent the analysis of contaminants in each organ requires not only Progress on Knowledge of the Behavior of Contamination in large individuals and extra treatments for the organs involved, Sediments and Their Toxicity to Aquatic Organisms, Workshop but also additional information about contaminants in an Proceedings, National Institute of Fisheries Science, Yokohama, organ.For example, it has been reported that contaminants 1994, pp. 142–154. included in the undigested remains in the intestinal tract of 6 E. I. Hamilton, Mar. Pollut. Bull., 1989, 20, 523. 7 D. J. H. Phillips and D. A. Segar, Mar. Pollut. Bull., 1986, 17, 10. mussels can clearly overestimate the accumulated concen- 8 D. Cossa, Oceanol.Acta, 1989, 12, 417. tration in viscera.35 However, we cannot yet estimate the 9 E. Gosling, in The Mussel Mytilus: Ecology, Physiology, Genetics degree of error caused by imprudent pretreatment of undiand Culture, ed. E. Gosling, Elsevier, Amsterdam, 1992, pp. 3–17. gested remains in the intestinal tract. It can be concluded from 10 K. J. M. Kramer, in Quality Assurance in Environmental these results that a composite sample of whole soft tissue from Monitoring—Sampling and Sample Pretreatment, ed.Ph. 30 blue mussels of 3–5 cm in shell length, collected from the Quevauviller, VCH Press, Berlin, 1995, pp. 179–211. 11 A. M. Caricchia, S. Chiavarini, C. Cremisini, R. Morabito and R. infralittoral zone at a time prior to the spawning season, is Scerbo, Anal. Chim.Acta, 1994, 286, 329. essential to reduce the variability between individuals under 12 J. L. Gomez-Ariza, E. Morales, R. Beltran, I. Giraldez and M. natural conditions and to obtain reproducible analytical values Ruiz-Benitez, Analyst, 1995, 120, 1171. at each sampling site. 13 S. Ohhira and H. Matsui, J. Chromatogr. B, 1990, 525, 105. The present study also revealed some variability due to the 14 H.Hraino and M. Fukushima, Anal. Chim. Acta, 1992, 264, 91. ecological aspects of mussels when used as exposure indicators 15 J. A. Sta�b and U. A. Th. Brinkman, Appl. Organomet. Chem., 1994, 8, 577. for organotin contamination. The multiple weight of the 16 Y. K. Chau, F. Yang and M. Brown, Anal. Chim. Acta, 1997, variability associated with the sampling date, sex diVerence, 338, 51.individual size and habitat height was estimated using a 17 M. Abalos, J. Bayona, R. Compan�o� , M. Granados, C. Leal and statistical method36 for the sum of each variability. If the M. Prat, J. Chromatogr. A, 1997, 788, 1. sampling strategy was not optimized against these ecological 18 M. Yoo and T. Kajiwara, Marine Fouling, 1983, 4, 11 (in parameters of mussel samples, the weight of variability Japanese). 19 T. Kajiwara, Y. Ura and N. Ito, Fisheries Sci., 1978, 44, 949 included in one analysed value was estimated to be (in Japanese). ±11.9–50.1% for TBT and 10.5–133.8% for TPT. The weight 20 N. P. Wilkins, K. Fujino and E. M. Gosling, Biol. J. Linn. Soc., of variability included in one analysed value using our optim- 1983, 20, 365.ized sampling strategy was estimated to be 16.2% for TBT 21 K. D. Daskalakis, Mar. Pollut. Bull., 1996, 32, 794. and 9.1% for TPT of the analysed value against the variability 22 R. B. Laughlin and W. French, Environ. Toxicol. Chem., 1988, due to analytical error only. We will be able to detect a 7, 1021. 23 R. F. Lee, Mar. Environ. Res., 1991, 32, 29. statistically significant diVerence of more than 20.6% for TBT 24 R.F. Lee, Mar. Ecol.-Prog. Ser., 1988, 46, 33. and 14.1% for TPT among the values of tissue concentration. 25 J. Capuzzo, J. Farrington, P. Rantamaki, C. CliVord, B. Judging from our study, the advantages of the use of blue Lancaster, D. Leavitt and X. Jia, Mar. Environ. Res., 1989, 28, mussels as exposure indicators may be summarized as follows: 489. (i) they permit the detection of contaminants that are present 26 H. Hummel, J. UniOudeGroeneveld, J. Nieuwenhuize, J. Liere, in ultra-trace levels or have no detectable concentration in R. Bogaards and L. Wolf, Sci. Total Environ., 1990, 92, 155. 27 P. B. Lobel and D. A. Wright, Mar. Pollut. Bull., 1982, 13, 320. seawater; (ii) they allow biological or ecological identification 28 A. J. S. Hawkins, E. Navarro and J. Iglesias, Mar. Biol., 1990, of potentially harmful chemicals in the environment; (iii) they 105, 197. can provide a rough approximation of chemical contamination 29 N. S. Fisher, J. Teyssie�, S. Fowler and W. Wang, Environ. Sci. and temporal and spatial comparisons of such contamination; Technol., 1996, 30, 3232. and (iv) they are stable environmental samples for storage and 30 K. Iida and M. Ogura, Annual Report of Kanagawa Institute of retrospective analysis.11 The limitations are summarized as Environmental Science, Kanagawa Institute of Environmental Science, Yokosuka, 1993, vol. 16, p. 45 (in Japanese). follows: (i) the dependence on the natural distribution of J. Environ. Monit., 1999, 1, 243–250 24931 A. L. Mallet and C. E. Carver, J. Exp. Mar. Biol. Ecol., 1993, ORCA 71, ed. G. Lauenstein and A. Cantillo, National Oceanic 170, 89. and Atmospheric Administration, Maryland, 1993. 32 D. M. Lowe, P. N. Salkeld and M. R. Carr, J. Mar. Biol. Assoc. 35 P. Lobel, S. Belkhode, S. Jackson and H. Longerich, Mar. UK, 1994, 74, 225. Environ. Res., 1991, 31, 163. 33 Y. Koshikawa, Y. Serizawa and A. Otsuki, IEEE Oceans 36 G. Christian, Analytical Chemistry, Wiley, New York, 4th edn., Proceedings, 1991, 2 (October 1–3), 1049. 1986, Section 3.8. 34 NOAA Technical Memorandum: Sampling and Analytical Methods of the National Status and Trends Program Benthic Surveillance and Mussel Watch Projects Analytical Protocols 1984–1992, NOS Paper 8/09669D 250 J. Environ. Monit., 1999, 1, 243&n