Table 4 — Average relative standard deviation and minimum and 



maximum values calculated from 1 12 (1988 through 1990 

 data) replicate treatment sets of three plots 





Relative stand deviation 



Average 



Minimum Maximum 







- - Percent 



S (Average species richness) 



10.2 



29.3 



H' (Shannon-Weaver index) 



6.1 



.5 26.4 



Table 5— Mean Shannon-Weaver diversity index (H' as log base 10) by herbicide 

 for all sites and times of application 



Pretreatment 1 year postspray 2 years postspray 



1988 1989 (1990) (1991) 





warm^ 



cool 



Cool 



Warm 



Cool 



Warm 



Check 



1.126 



1.365 



1.363 



1.257 



1.345 



1.281 



Tordon 



1.082 



1.345 



1.262 



1.164 



1.319 



1.253 



Stinger 



1.077 



1.349 



1.340 



1.240 



1.346 



1.307 



Curtail 



1.090 



1.344 



1.311 



1.198 



1.341 



1.290 



Mn this table warm/cool (season) refers to when the community data were collected: 

 cool = May and June; warm = July and August. 



Table 6 — Average number of species by herbicide for all sites and times of 

 application 



Pretreatment 1 year postspray 2 years postspray 





1988 

 warm^ 



1989 

 cool 





(1990) 





(1991) 



Cool 



Warm 



Cool 



Warm 



Check 



25.6 



38.8 



40.2 



34.8 



37.7 



33.8 



Tordon 



24.7 



38.4 



34.4 



28.7 



35.8 



32.9 



Stinger 



24.5 



39.2 



39.0 



33.8 



38.0 



35.0 



Curtail 



25.8 



38.7 



37.6 



31.5 



37.8 



34.4 



Mn this table warm/cool (season) refers to when the community data were collected: 

 cool = May and June; warm = July and August. 



herbicides on diversity was small and transitory 

 (figs. 1, 2). Tordon-treated plots had the lowest av- 

 erage diversity 1 and 2 years postspray, Curtail was 

 intermediately lower, and Stinger plots were similar 

 to the check plots. The 1-year postspray differences 

 between treatments decreased in the second-year 

 postspray. The measured average number of species 

 and the Shannon- We aver index were higher for the 

 Stinger and Curtail plots than for the checks by the 

 second-year postspray. 



Analytic precision and accuracy of the herbicide 

 residues in the soil samples was good. Recovery of 

 known additions to prespray collected soil blanks 

 from the research sites averaged 88 percent {n = 19). 



Duplicate analyses of split field samples had an av- 

 erage relative standard deviation of 6.8 percent 

 (n = 32). We detected no herbicides in the reagent 

 blanks (n = 18). The method detection limit (MDL) 

 was 10 parts per bilhon (ppb). Instrumental re- 

 sponses below MDL are reported as trace. 



Herbicide residues in the soil decUned rapidly with 

 time and depth (table 7). Although it had the high- 

 est svuface layer concentrations at day 30 because of 

 its higher initial appUcation, 2,4-D nevertheless de- 

 graded most quickly. The maximum concentration 

 of 2,4-D below 25 cm at day 30 was 14 ppb. Only 

 site 1 early treatment stiU had trace (<10 ppb) de- 

 tectable 2,4-D below 5 cm after 1 year. As expected, 



4 



