122 



ANOVA for the untreated control (Table 31) was run to determine 

 if there was a difference between the controls. The analysis determined 

 a significant difference at the .05 level. The two untreated control 

 populations were different, though there were no treatment (T) x week 

 (W) interactions. The lack of a TxW interaction meant that, even though 

 the controls continued to have different population levels, this level 

 was maintained from week to week and did not vary significantly between 

 controls. The paralleling values of the control data were averaged 

 for comparison to the treatments. The averaged control data with each 

 treatment are given to demonstrate the response of each chemical (Figures 

 35-40). 



The chemical standard, chlorobenzylate, was included in the ANOVA 

 of the treatments (Table 32). The blocks (rows) were found to be 

 highly significant. For this reason, blocks and replicates are em- 

 ployed. The treatments were significantly different at the .05 level. 

 This demonstrated the fact that there was a difference between the 

 effects the chemicals had on the rust mite. The treatment x week 

 interaction, TxW, failed to show any significance at the .01 or .05 levels. 

 This demonstrated the test's inability to detect any differences be- 

 tween the treatments as the weeks went on. That is to say, the treat- 

 ments were different but this difference did not significantly vary 

 from week to week in the 1976 study. 



The Duncan's new multiple-range test (Table 33) was tested across 

 the weeks instead of by week (as in 1975) due to the inability to demon- 

 strate any treatments by week interaction. Banex (Treatment 6) at 

 1.6% concentration, demonstrated the least control of the products 



