66 



tion, as in the mature plant. It is, however, to be ob- 

 served that the number of scale-like leaves is far from con- 

 stant ; there maybe developed merely two or three, as shown 

 in Fig. 33. The same course in regard to the development of 

 scale-like leaves succeeded by normal ones is to be noted in 

 the full-grown plant, as shown in Fig. 35, where six scale-like 

 leaves precede the two normal ones and the flower. The 

 shape of the full-grown leaf is palmately lobed, but it shows a 

 great variation in regard to the number of lobes (three or 

 nine) and also in regard to the depth of the sinuses. But it ^ 



is a constant character, that, not only the normal leaves of the 

 young specimens, but alsb the first or second ones, belonging 

 to the lateral branches of the main subterranean stem, the 

 large rhizome, are similar in shape, always heart or kidney- 

 shaped, without lobation. I did not observe in any case that 

 the normal leaves on the lateral branches were not preceded 

 by scale-like ones, and Fig. 36, which represents a lateral 

 shoot, shows the presence of seven scale-like leaves, after 

 which come two normal ones, Li and L^. The full-grown * 



rhizome (Fig. 35) is rather long, cylindrical, and dies off 

 gradually at its posterior end. Most of the roots are strong, 

 sparingly branched above, and very long, especially those 

 proceeding from the inferior part of the rhizome. In regard 

 to the whole growth of the rhizome, this has been shown to 

 be sympodial by Mr. Foerste in his interesting paper, 

 ** Notes on Sanguijiaria Canadensis'''^ where he has also 

 mentioned the singular case of the occurrence of a two- 

 flowered scape. 



* Bulletin of the Torrey Bot. Club, Vol. XIV, No. 4, 1887. 



r 



r ' 



yiOLACE/E. 



Viola pabnata, var. aiadlata. 



The germinating plantlet of this species (Plate VIII, Fig. \ 



37) shows the long-petioled, ovate cotyledons, the distinct 

 hypocotyl (C), the imbranched primary root (R) and the two 

 secondary ones (r^), w^hich have been developed at the same 

 time. There is a great difference between this early state and 



\ 



