EST THE TROPICS : HI. 177 



The results here shortly described confirm those of Tschermak 

 in a qualitative manner; quantitatively the results were 

 uncertain owing to bad conditions, which only a few plants 

 survived. 



A considerable number of crosses were made at Peradeniya 

 between strains of beans which' differed chiefly in the colour 

 of the testa. Unfortunately the greater number of the pro- 

 geny of these crosses were lost owing to drought and other 

 causes. The number of plants which survived was in no case 

 sufficient to afford complete evidence of the manner of inheri- 

 tance of the characters examined, and the phenomena ob- 

 served will only be very briefly indicated. 



Besides the case of atavism described on page 340, the only 

 example of a bean cross which survived more than two gener- 

 ations was one between " Golden Wax Butter " and " Ne Plus 

 Ultra ; ' ' the former of these had yellow pods and testas of a 

 dark blackish purple colour, whilst Ne Plus Ultra had green 

 pods and dark yellow seeds. 



The seeds of the cross bred plants (Fi) resembled those of 

 Golden Wax Butter, or were perhaps of a slightly darker tint, 

 but the pods were green. And from the evidence of later 

 generations it was clear that the green pod colour behaved as a 

 simple dominant to yellow. 



As regards the characters of the testas, a great variety of 

 types were obtained in F 2 and F 3 . The number of plants 

 obtained was small, but the facts appear to be as follows. The 

 self -fertilized heterozygote gave rise to two principal groups 

 of plants — a group (A) with seeds showing various shades of 

 purple, including sub-groups with reddish purple, bluish 

 purple, and chocolate testas — the whole group being pro- 

 bably three times as numerous as the second group (B^ 

 which included different shades of yellow, mottled yellow 

 and white, and pure white seeds. On further self-fertilization 

 members of group B never gave rise to any form of group A, 

 but apparently any of the A types might produce some form 

 of B. 



