63 



over a curved red line running downward from the top of the 

 little lobe (formed by the notch) and ending in a shallow 

 groove. In both flowers there is a crack to be seen just between 

 the sepals in question. From this we gather that a portion of 

 a divided petal adheres to each of the margins. 



The same phenomenon is seen in a tetramerous specimen. 

 Also in this case there is a petal short. In this flower the 

 incompleteness of the tetramery goes still a step further since 

 the two sepals flanking the open space correspond to only one 

 cell of the three-celled ovary. 



Sisyrmchium latifolium Sw. 



Habitat Japan. 



Coll. July 1895. 



A normal flower with four stigmas. 



MARANTACEAE. 



Canna spec, et varietates diversae. 



Coll. Aug. 1894, Sept. 1894, Oct. 1894, June 1895, July 1895, 

 Nov. 1895, June 1898. 



A. FLOWERS. 



In order to facilitate a survey of the numerous aberrations 

 observed in the flowers of Canna and at the same time to 

 make clear the significance of many of these aberrations it is 

 necessary first to give a brief description of the flower and to 

 state that we follow Eichler's hypothesis as expressed by this 

 authority in fig. 5 '). The flower has 3 very short sepals, 3 much 

 longer petals, next 3 or 4 petaloid organs, 1 stamen and 1 pistil 

 of which the style is petaloid and somewhat slanting and stunted. 

 Of the 3—4 petaloid parts the biggest is called labeUum, the 

 others cdae. Eichler indicates the alae with a and /?, if there 

 are two; the third is by him called y and in his diagram takes 



1) As is well known, Eichlek in his sBluthendiagramme" I, p, 174, offers two 

 different explanations of the flower of Canna, without expressing preference for either. 

 We intend to account further on for our preferring his flrst explanation, as we have 

 already done in Ned. Kraidk. Archief 1902. Fig. 6 represents his second explanation. 



