Table3. Results of sign test applied to mean chemical concentrations at the 141 sites sampled 

 in at least four of the years between 1986 and 1990 s . 



a Comparisons made on a year-to-year basis and between 1986 and 1990. A statistically 

 significant (0.05 level) proportion of changes in the increasing (Inc) or decreasing (Dec) 

 direction are indicated, as are cases with no significant direction of change (-). Since it was 

 not routinely measured in earlier years, tBt comparisons have been made for the 1 49 sites 

 common to 1989 and 1990. 



b tPAH data for 1 986 were not used. Geometric mean for that year was 430 ng/g which is 

 1 .4 times the next highest annual geometric mean. Its high value may have been due to 

 the analytical method used in that year for samples from the East and West Coasts. The 

 last comparison for tPAH is an 87-90 comparison. 



given the implementation of pollution 

 controls and changes in industrial prac- 

 tices make any difference, decreases 

 are expected. In fact, the overall con- 

 clusions from Table 3 are: ( I ) usually, 

 between-year differences were not statis- 

 tically in one direction; (2) when a direc- 

 tion was indicated, it was generally a de- 

 crease: and (3 ) when 1 990 is compared with 

 1986, the only changes were decreases. 



The sign test has been applied to the 

 directions of concentration differences. 

 Because no account has been taken of 

 the amount of change, small and large 

 differences have had equal bearing on 

 results. To provide a sense of the mag- 

 nitude of annual changes, the annual 

 geometric means of chemical concen- 

 trations among all 141 sites sampled in 

 four or five years are shown in Figures 



16 



