8 CIRCULAR 9 5 2, U. S. DEPARTMENT OF AGRICULTURE 



provide no definite evidence that observed off -flavors can be attributed 

 to the production of peanuts in rotation with BHC-treated cotton. 

 Among the 65 treated commercial samples that followed BHC-treated 

 cotton, Panel A scored only 4 significantly lower on flavor than their 

 comparable controls, while 3 were scored higher. Separate analyses 

 were made of scores from sessions in which the control was included 

 as an unknown; for this reason differences between sample and control 

 scores for some sample means within such groups were found to be 

 statistically significant whereas comparable or even greater differences 

 were not significant in the overall analysis. In the overall analysis, 

 no untreated sample was scored significantly lower than its control 

 and no significant correlations between palat ability scores and reported 

 insecticide applications were observed. 



One of the lowest scoring samples (table 2) was interplanted in 1950 

 with BHC-treated cotton and was therefore exposed to direct contact 

 with the insecticide. The BHC content of this sample, 0.49 p. p. m., 

 was almost four times as high as that of any of the other 1950 com- 

 mercial samples. 



Trends in the results of the scoring by Panel B (tables 1, 2, 3) 

 provide a picture generally comparable to that of Panel A. Of the 

 treated commercial samples, 5 were scored significantly lower and 2 

 higher than their controls. Two untreated samples were also scored 

 significantly low but off-flavor in one of these was attributed to im- 

 proper curing (table 3). Two of the 3 samples that followed cotton 

 receiving the heaviest reported insecticide treatments (table 1) were 

 scored significantlv low and these were among the 15 samples showing 

 a BHC content of 0.10 to 0.13 p. p. m. 



Among the samples that were scored low on flavor, the 2 panels 

 agreed on only 1 (table 3). This was also one of the samples having 

 a relatively high BHC content (0.14 p. p. m.). Aside from the above 

 noted samples, however, there was no apparent correlation between 

 flavor scores and chemically determined BHC content of the 1950 

 samples. 



A general picture of the results is provided by the scatter diagrams, 

 figures 1 and 2. These figures show the differences between panel 

 scores for the commercial samples and those for their comparable 

 variety- controls. The distribution of score differences by Panel A 

 (fig. 1) shows no apparent correlation between reported BHC dosages 

 and panel scores. The number of scores for the treated commercial 

 samples that fall above and below scores for the comparable control 

 samples is approximately equal. A similar distribution of score 

 differences is shown for the untreated commercial samples. The 

 same general pattern is presented by Panel B scores (fig. 2). 



The only valid conclusion that may be derived from these data is 

 that of the commercial samples the general average scores of both the 

 "treated" and the "untreated" samples did not differ significantly 

 from the average scores of the untreated controls from experiment 

 station plots. The few observed instances of significant differences 

 between single samples and their appropriate experiment station 

 controls were not definitely correlated with dosage of BHC. In 

 connection with the commercial samples it should be emphasized 

 that: The information on insecticide treatments used for cotton 

 crops was based on recollection of growers as to times, frequency, and 

 quantities of application, rather than on written records; methods of 



