-29- 



Table III is somewhat like table II, but it presents the 

 appacnptly poorest solutions for shoot grov/th instead of the appa- 

 rently best ones. '-The soliitions shovrn in this list are those 

 whose grov;th values are among the four lov^est of their 

 respective series, for both tests, and for which the growth 



values for the two like tests show differences _o-f .11 mm^w 

 js^ le&i=i« Otherwise, the notation follows the scheme of 

 table II* 



