io6 BOTANICAL GAZETTE [FEBRUARY 



less than nine branches, the average number of branches in the latter 

 being about 15. 



During the same season, 81 offspring of an unbranched plant 



(04108) which had also been left unguarded showed 72 unbranched 

 and 9 branched. This result would find a ready explanation in case 

 branching is a Mendelian unit-character, simply as the usual result 

 of vicinism. On this assumption the proportion of branched off- 

 spring in this family indicates that the pistil-parent (04108) received 

 about one-fifth of its fertilizations from the branched specimen 



(04109) which was growing near it, and the other four-fifths from 

 the no unbranched plants, several of which were growing equally 

 near. 



In 1906 the Helianthus cultures were started several weeks too 

 early (April 2) and became tall and slender from long crowding and 

 insufficient illumination before they could be set into the garden. 

 Some of these began to bloom shortly after they were transplanted, 

 and the branching habit was very much deranged. Two families, 

 (05149 and 05150) raised from branched parents having had the 

 pollination fully controlled, resulted in 54 branched and 2 nearly 

 unbranched in the one case, and 68 branched and 2 unbranched in 

 th other, but as the whole branching system was considerably modi- 

 fied the classification could not be made with security. The expecta- 

 tion in these cases, since both parents were DR, would have been 

 branched and unbranched approaching the ratio 3:1. The results 

 secured indicate either that we are not dealing with a Mendelian char- 

 acter or that the conditions of the culture caused an excessive develop- 

 ment of branches in normally unbranched individuals. The latter 

 suggestion is strengthened by a small family (05145) reared from fully 

 guarded unbranched parents from the same hybrid family to which 

 the branched parents of the last-mentioned two families belonged. 

 Both parents should be in this case extracted recessives (RR), and 

 all of the offspring should belong to the unbranched class. Only six 

 offspring were produced, and all showed some branching, ranging 

 from one strong branch below the middle to twelve branches from the 

 middle to the top of the stem, the latter type of branching being char- 

 acteristic of the branched form. The results in 1906 seem therefore 

 to oppose distinctly the idea that the branching character is Men- 



