Inheritance of the heptandrarform. of Digitalis purpurea L. 263 



from these hand-pollinations, and to insure against the loss of the 

 strain, seeds were also collected from specimens whose flowers had 

 been unguarded against promiscuous cross-pollination. In 1908 sowings 

 were made from all of the attempted crosses and also from these 

 unguarded seeds, but only the latter resulted in germinations. Of 

 the plants so produced (family 0745) 184 came to bloom in 1909, 

 154 having white flowers of the heptandra-type in various grades of 

 development, thus agreeing with their seed-parent, and 30 with 

 flowers of the normal type, 18 with purple flowers and 12 with 

 white flowers. 



Of the heptandra plants in this family 10 produced no open 

 flowers because of damage from thrips; 3 had corollas of nearly nor- 

 mal form but with the lower lip margined with half-anthers, or par- 

 tially split into three segments, each bearing an anther with divergent 

 lobes; 95 had a well developed upper lip, truncate or but little 

 divided, and seven stamens; 40 had the upper lip more or less 

 strongly cleft or laciniate, with two of its segments frequently bearing 

 half-anthers, or rarely with complete anthers; 4 had the upper lip 

 almost entirely absent in most of the flowers; one had the corollas 

 split into four segments, the lateral and lower segments each being 

 terminated with rudimentary anthers, thus standing intermediate be- 

 tween the class with the least development of the 7z^/>to^rtf-characters 

 and the modal condition in which the lower lip is replaced by three 

 perfect stamens, while the upper lip remains but little modified. 

 These categories are not quite the equivalent of Miss SAUNDERS'S 

 classes, for the plants here recorded as having the upper lip laciniate, 

 doubtless represent the transitional stages from the simple 7-stamened 

 to the 9-stamened classes in her grouping, and while many of them 

 would probably have been placed in her "heptandra-imm proper (i. e. 

 with a large number of flowers having a petaloid upper lip", some 

 of the more extreme cases of laciniated upper lips might have been 

 classified without much violence with her "extreme heptandra-ivrm" . 

 It must be borne in mind that all of these classes are artificial, and 

 are connected by intermediate forms; but the records of the fluctu- 

 ations among my plants correspond so closely with those of Miss 

 SAUNDERS, that it may be fairly inferred that the range of fluctua- 

 tion in the two cases has been essentially the same. 



No doubt was entertained that the normal plants which appeared 



in this family were produced as a result of pollination by insects 



; coming from normal plants in the adjoining garden. Such normal 



! 



