a wet- weight basis) above specified lev- 

 els of mercury and several chlorinated 

 hydrocarbons. With one exception, none 

 of the mollusks collected in the NS&T 

 Program have had concentrations (ad- 

 justed to a wet-weight basis) that exceed 

 those levels. The exception is the tPCB 

 concentration in mussels collected at the 

 Angelica Rock site in Buzzards Bay, 

 MA (near New Bedford Harbor) in 1 989, 

 but not in prior years nor in 1990. 



TEMPORAL TRENDS 



Mussels and oysters are collected by the 

 NS&T Program to monitor trends in 

 chemical concentrations over time. 

 Chemical concentrations in mussels and 

 oysters are determined by the extent to 

 which the organisms accumulate chemi- 

 cals from the food they filter from their 

 surrounding water and from the water 

 itself. When chemical concentrations 

 increase or decrease in their surround- 

 ings, the organisms are capable of in- 

 creasing or decreasing the correspond- 

 ing concentrations in their tissues (see 

 forexample,Roesijadietal., 1984;Pruell 

 etal., 1987). This, and the fact that they 

 are immobile, make them ideal for moni- 

 toring changes in chemical concentra- 

 tions at fixed sites. 



Temporal trends can be defined through 

 two approaches. First, a trend exists 

 when a year-to-year change in the same 

 direction occurs at the vast majority of 

 sites. Secondly, there can be a statisti- 

 cally meaningful relationship between 

 concentration and time. The butyltin 



data provide an example of the first 

 kind of trend. Detectable concentra- 

 tions were found at 149 sites in both 



1989 and 1990. Therefore, in 149 cases 

 butyltin concentrations could have in- 

 creased or decreased. While no impor- 

 tance can be attached to the direction of 

 change at any one site, it is statistically 

 significant that butyltin was lower at 

 103 sites in 1990. The probability of 

 flipping a coin 149 times and getting 

 103 "heads" is very small. Similarly, 

 there is practically no chance that 103 

 decreases in butyltin between 1 989 and 



1990 was a random event. Strictly 

 speaking, the chance of that result oc- 

 curring randomly is less than 1 in 1 ,000. 

 Here, and throughout this report, re- 

 sults that are likely to occur randomly 

 less than 5 times in 100 (0.05 level of 

 significance) will be considered impor- 

 tant. 



Five years of data have been collected 

 for the other chemicals. While the 

 number of sites has increased over the 

 years, comparisons between years will 

 be limited to the 141 sites that were 

 sampled in at least four of the five 

 years. Applying the same test just 

 applied to the tBT data (called the sign 

 test) yields the results summarized in 

 Table 3. There is no need to consider 

 oysters and mussels separately for any 

 chemical, since we are looking for dif- 

 ferences at individual sites. Between 

 1 989 and 1 990, when butyltin was found 

 to decrease, tPCB also decreased, and 

 four chemicals — silver, arsenic, lead, 

 and zinc — increased. Despite that, 



15 



