126 LUTHER BURBANK 



And wben we advance from five pairs of unit 

 characters to ten or twelve, as wr have already 

 aeen that we must do in the cas< r cherr\ 



the matter beeomet almost infinitely codm 

 Aa we increase the number of qualities un<i< 

 consideration, the number of possible comhina 

 tions among them increases at an alarn .. 

 geometrical ratio. 



It appears that wliereas there is an ercn 

 chance, when only a single pair of qualities was 

 in question, of producing one offspring like each 

 parent in each group of four; and whereas there 

 is the same even chance of producing one off- 

 spring like each parent in every group of 256 

 individuals when four pairs of unit characters 

 are in question — when we have to deal with ten 

 pairs of unit characters the possible arrangt 

 ments have become ao bewildering and complc 

 that there is even chance of producing a single 

 offspring like each grandparent only iti each 

 group of more than a miUion progeny! 



Quantity PaoDuciioN Xecessast 



Such a computation, as made in accordance 

 with the Mendelian formula, in itself serves to 

 supply a ready answer to those Mendelians who 

 have questioned the necessity of making experi- 

 ments on the elaborate scale that I have all along 



