36 Journal of Agricultural Research voi. xix, no. i 



It may seem that, except for the character chosen, the values for the 

 most part depart little from the mean values. For example, under 

 total leaves the most maize-like plant which had 13 leaves was partic- 

 ularly maize-like in no other character. It was even below the average 

 in number of alicoles in the best spike and had almost the maximum 

 number of suckers. On the other hand, the plant with the greatest 

 number of leaves had also the greatest number of male branches but was 

 decidedly maize-like with respect to number of suckers and number of 

 alicoles. 



CONCLUvSIONS 



The genetic relations of the principal characters of maize and teosinte 

 were investigated in a cross between a small variety of pop com and 

 Florida teosinte, a large forage grass generically distinct from maize. 

 The Fj plants showed characters which, for the most part, were interme- 

 diate between those of the parents. 



The F2 plants were also intermediate, with a greatly extended range 

 of variation. Thirty-three of the characters that differentiate the parents 

 were chosen and recorded for each of the 127 Fj plants. The distribution 

 of these characters with one or two exceptions showed little or no evidence 

 of alternative or Mendelian inheritance. 



With respect to the individual characters, the extreme variants 

 approached, and in some instance exceeded, those of the parents; but 

 none of the plants possessed any large number of the characters of either 

 maize or teosinte. 



The results showed the greatest freedom of recombination. All com- 

 binations of characters appeared that might reasonably be expected 

 with so limited a number of individuals. There were many instances 

 of coherence or partial coupling, but there was an almost equal number 

 of instances where characters derived from different parents showed a 

 tendency to combine more frequently than would be expected as the result 

 of chance. In such a complicated series it was found impossible, how- 

 ever, to distinguish primary from secondary correlations. 



While there appeared to be no incompatible combinations, there were, 

 on the other hand, no completely independent characters. Every 

 character recorded showed significant correlation with one or more 

 other characters; and these in turn were correlated with still others, with 

 the result that all the characters were interrelated and formed a single 

 group. It is possible, in fact, to arrange all the characters in such a 

 way that they form a single group in which there is no coefficient of 

 correlation lower than ±0.31. 



The nearest approach to Mendelian inheritance was shown by the 

 arrangement of the spikelets in the pistillate inflorescence (fig. 18, 19, 

 20). In maize the female spikelets are borne in pairs (double female 

 alicoles) ; in teosinte the female spikelets are borne singly (single female 



