4a to 4g. Here, since oyster sites and 

 mussel sites are both part of the national 

 mean, it has been necessary, again, to 

 separate the data between species for 

 silver, copper, zinc, chromium, and lead. 



Decreases or increases identified in Table 

 3 also appear in Figure 4. The important 

 conclusion from Figure 4 is that changes 

 are small. For both elements and or- 

 ganic compounds, annual shifts in geo- 

 metric means are generally less than 20 

 percent. 



The sign test is useful when there are 

 many comparisons to be made between 

 any two years; it exploits the fact that 

 there are many Mussel Watch sites. The 



second method of trend detection iden- 

 tifies connections between concentra- 

 tion and time at individual sites and 

 depends on the number of years of data. 

 While plots will not be shown of annual 

 mean concentrations of each chemical 

 at each Mussel Watch site, this tech- 

 nique can be illustrated with two long- 

 term data sets. Figure 5a shows the 

 annual average carbon dioxide concen- 

 tration in the atmosphere at the Mauna 

 Loa Observatory in Hawaii for each 

 year from 1959 through 1985 (Keeling 

 andBoden, 1986). Figure 5b represents 

 the annual mean temperature in the 

 Northern Hemisphere from 1 85 1 to 1 984 

 relative to the average temperature over the 

 years 1951 through 1970 (Jones etal., 1986). 





CD 

 O 



a 

 o 

 u 



a 



TO 

 CD 



H 

 o 



CD 



o 



CD 



0.6 

 0.5 



■o 



i>0.3 



CD 



0.2 



0. 1 

 



Ag (mussel) 





•^ .^7. i. { .*» 



Hg 





i 



i 



85 86 87 88 89 90 



year 



9 1 



Figure 4a. Annual geometric mean concentrations of selected chemicals in mollusks. 



17 



