460 



-18- 



equivaleot to soaa of McKcan«ll't eluateri whereat others have no rc4l 

 equivalent. 1h« Cvo sets of clusters are coopared la Table 5 belov. 



TABLE 5 



I 



Deciding which of McKennell'a cloiters came closest to each of the 

 WHEAT clusters involved plotting and comparing the cluster profiles. 

 These are reproduced at the end of the report.. The more complex cluster 

 solution agreed for Project WHEAT appears to have Resulted in the 

 sub-division of two of McKennell's clusters (III and^I) , and the plotted 

 profiles show a fairly good match between McKeonell HI and^the average 

 of WHEAT 3 and 9, likewise between McKennell VI and the avera^^of 

 WHEAT 8 and 11. On the other hand WHEAT 12 seems to contain elemints 



09 



c 

 I 



c 



N 

 t 

 O 



o 



