20 THE ROYAL SOCIETY OF CANADA 



years, for it was found that whenever the six rows of kernels were 

 fully developed the plant was homozygous in regard to rows, but if 

 even a very few kernels were lacking at the base of the head such 

 plants were heterozygous. The six-row condition was nearly, but 

 never completely, dominant. These facts are of special interest in 

 view of the generally accepted idea that the two-row condition is 

 essentially dominant over the six-row. 



With this new information, let us further study our types so as 

 to clearly understand what inheritance ratios may be expected. 



Type I contains: 



(A) Two-row plants homozygous for hoods. These plants breed 

 true to the same type. 



(B) Two-row plants in which awns are recessive. These plants 

 produce types 1(A), 1(B) and III. 



Type II contains: 



(A) True six-row plants, homozygous for hoods. These plants 

 breed true, of course. 



(B) True six-row plants in which awns are recessive. These 

 plants produce types 11(A), 11(B) and IV(A). 



(C) Pseudo six-row plants, homozygous for hoods. These plants 

 produce types 1(A), 11(A) alnd 11(C). 



(D) Pseudo six-row plants in which awns are recessive. These 

 plants give rise to all the types, namely, 1(A), 1(B), 11(A), 11(B), 

 TI(C), 11(D), III, IV(A) and IV(B). 



Type III is simple and always breeds true. 

 Type IV contains: 



(A) The true Arlington type (six-row with poorly developed 

 awns on the median rows only). These plants breed true. 



(B) The pseudo Arlington type (the four lateral rows having a 

 few gaps, usually at the base of the head). These plants produce 

 types III, IV(A) and IV(B). 



We are now in a position to reconsider the numerical ratios 

 recorded above. In the light of the further information available we 

 see that the ratios between the four types should be 3:9:1:3 when 

 the plants of the second generation are counted. Taking the number 

 of plants of type III as our unit (127) we should expect 381 plants 

 of type I, instead of 420 observed, 381 plants of type IV, instead of 

 477 observed, and 1143 plants of type II, instead of 997 observed. 

 The discrepancies are very large. It must be noted, however, that 

 types I and II were not always reasonably easy to distinguish, for 

 some plants had about half of the kernels developed in the lateral 

 rows and could not be classified with certainty. Probably too many 



